* WIP
* WIP
* WIP
* Testing Trellis quantization
Using 12 bits per 8 weights I get a better rmse than
iq2_xxs. I still need to see how quantizing the group-of-8
scales will affect accuracy. By AVX2 SIMDifying the search
for the best code, LLaMA-3.1-8B gets quantized in 130 seconds
on the Ryzen-7950X CPU - sluggish but still acceptable.
* Testing Trellis quantization: 4-bit quantized block scales
rmse increases by just 3%, so this is beating iq2_xss in terms
of rmse at the same 2.0625 bpw.
* Testing Trellis quantization: playing with scales and generators
* iq2_kt: quantize / dequantize
I now see that I was comparing apples to oranges:
iq2_xxs was using a weight of sigma^2/4 + x^2, while
the Trellis approach wasn't (weight = 1). Once I use the same weight,
iq2_kt is actually slightly worse than iq2_xxs in terms
of rmse, so does not look promising at this point.
Also, once each group of 8 Trellis values no longer has a
constant sum(q^2) that we can precompute, quantization
becomes significantly slower (476 seconds for LLaMA-3.1-8B).
* iq2_kt: CUDA dequantize
so we can run perplexity calcs.
As already indicated by rmse, the 2-bit trellis approach is
quite a bit worse than iq2_xxs.
* WIP
* WIP
* WIP - try larger blocks
With blocks of 32 and 16 bits per groups of 8 the brute force
seach becomes prohibitive in terms of CPU time (30+ minutes
for 8B LLaMA after SIMDifying with AVX2). The trick is to
group the points in clusters, find the nearest cluster,
and only search within the cluster.
* iq2_kt - this is better
Using blocks of 32 and 16 bits per group of 8 weights
it beats iq2_xxs in terms of PPL by a significant margin.
It is 0.0625 bpw larger, but even if we go to 15 bits per
group od 8 (so 0.0625 bpw less than iq2_xxs), PPL is still
lower.
* iq2_kt - even better
Re-quantize after determining block scales
(at the epxense of much longer quantization time).
* iq2_kt: CUDA dot product
Implemented as DMMV.
Very slow - just 81 t/s for LLaMA-3.1-8B.
Then again, Q2_K_S with forced to use DMMV only
gets 112 t/s vs 145 t/s via MMVQ. My memory is that
when the DMMV kernels were properly maintained/used,
DMMV was about on par with MMVQ for k-quants on my GPU.
* iq2_kt: very slightly faster CUDA dot product
* iq2_kt: f16 CUDA dot product
We arrive at 112 t/s.
* iq2_kt: faster f16 CUDA dot product
We arrive at 139 t/s (no FA), and 149 t/s (FA).
My RTX-4080 is ~20% slower than the RTX-6000 quoted in the
QTIP repository, so with FA (which I'm sure they also used)
we are at around ~180 t/s on their GPU, so almost matching
their performance.
* iq2_kt: faster f16 CUDA dot product
We arrive at 146 t/s (no FA), and 158 t/s (FA).
This is measured for LLaMA-3.1-8B with output.weight
left as f16.
* Minor
* Adding iq3_kt
3.125 bpw. So far does not look good on the PPL vs bpw plot.
* Forgotten change
* WIP
* WIP
* iq3_kt WIP: slowly improving
PPL(LLaMA-3.1-8B-Instruct, 8192) is now 6.8322, which is
starting to be competitive/slightly better than other quants.
* WIP
* iq3_kt WIP: slowly improving
PPL(LLaMA-3.1-8B-Instruct, 8192) is now 6.7892
* iq3_kt WIP: slowly improving
PPL(LLaMA-3.1-8B-Instruct, 8192) is now 6.7689 after shrinking
by 0.015 bpw by using iq4_k instead of q5_k for attn_v.
* iq3_kt WIP: speed up quantization
Nearly 60% improvement of quantization speed by having the
points nelonging to a cluster copied to contiguous memory
during initialization, and then accessed sequantially while
searching for the closest point. LLaMA-3.1-8B now gets
quantized in ~150 seconds on the Ryzen-5975WX.
* iq3_kt speed up quantization
Same trick as last commit applied to iq2_kt. Here we get
an even larger speedup: quantization time on the Ryzen-5975WX
for LLaMA-3.1-8B drops to 195 seconds from 375 seconds!
* iq3_kt: CUDA dot product
* iq2_kt: SOTA
We arrive at
PPL(LLaMA-3.1-8B-Instruct, 8192) = 9.2406
PPL(LLaMA-2-7B, 4096) = 6.4179
* iq2_kt: SOTA
We arrive at
PPL(LLaMA-3.1-8B-Instruct, 8192) = 9.1642
PPL(LLaMA-2-7B, 4096) = 6.3920
* Adding iq4_kt - not competitive at this point
* WIP
* WIP
* iq4_kt: CUDA dot product
* iq4_kt: minor tweaks
* iq2_kt: SOTA
We arrive at
PPL(LLaMA-3.1-8B-Instruct, 8192) = 9.1642
PPL(LLaMA-2-7B, 4096) = 6.3920
* iq2_kt: SOTA
We arrive at
PPL(LLaMA-3.1-8B-Instruct, 8192) = 9.0297
PPL(LLaMA-2-7B, 4096) = 6.3913
Ah, quantization is faster too. About 20% faster.
* iq3_kt: small improvements and faster quantization
* iq2_kt: SOTA
We arrive at
PPL(LLaMA-3.1-8B-Instruct, 8192) = 8.9627
PPL(LLaMA-2-7B, 4096) = 6.3825
Quantization is faster too: ~200 seconds for LLaMA-3.1-8B
on Ryzen-5975WX.
* iq3_kt: small progress
* WIP
* iq4_kt: go to 4.0 bpw
15 bits per group of 4, plus 8 bit scales ifor blocks of 32.
This gives a slightly better PPL than iq4_kss.
* iq4_kt: very slightly better
at the expense of much longer quantization time.
* iq4_kt: failed attemt to adjust CUDA dot product
It was working for 4.125 bpw. But after changing to 4.0 bpw
there is something wrong and I don't see the bug.
* DRY
* DRY
* iq4_kt: CUDA dot product works
* DRY
* Report actual bpw
* Minor tweaks
* Checkpoint
Go to groups of 8 for iq3_kt. 2 x 8 = 16 bits for the magnitude
plus 1 bpw for the sign. It goves a visible improvement in the
PPL vs bpw plot, but that comes at the expense of much longer
quantization time (7.5 minutes for LLaMA-3.1-8B on the Ryzen-5975WX).
I also notices that the 3INST generator is not actually generating a
Gaussian distribution. But going to a better generator means
readjusting all the hyper-parameters, so leaving it for later.
* WIP for IQ2_KT
* WIP - working basic iq2_kt
* still super slow (0.17t/s eval)
* flatten 3inst iters + avx2 (0.3t/s eval)
* iq3_kt (0.3t/s eval) and renames
* wip buggy iq4_KT
* fix (0.22t/s eval)
* naming and remove unused fn
* cleanup
* more cleanup
* delete unused and noncompiling mmvq functions
* Some performance tweaks
* Slighty faster iq2_kt
* port Trellis struct to iq3_kt, iq4_kt
* oops untracked files
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
gguf-split : improve --split and --merge logic (#9619)
* make sure params --split and --merge are not specified at same time
* update gguf-split params parse logic
* Update examples/gguf-split/gguf-split.cpp
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
---------
gguf-split : add basic checks (#9499)
* gguf-split : do not overwrite existing files when merging
* gguf-split : error when too many arguments are passed
Authored-by: slaren <slarengh@gmail.com>
* iq5_ks: basics
* iq5_ks: quantize
* iq5_ks: CUDA dequantize works
* iq5_ks: dot product works on CUDA
* iq5_ks: MMQ works
* iq5_ks: Zen4
* iq5_ks: AVX2
But is is not quite right, just like iq4_k, iq5_k, iq6_k, iq4_ks.
All these need fixing on AVX2.
* iq5_ks: NEON
* iq5_ks: Metal dequantize
* iq5_ks: Metal dot product
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* imatrix: collect layer influence statistics
* imatrix: collect layer influence statiscs also for the last layer
For the last layer we need to use the input for the output.weight
tensor. Last layer(s) tend(s) to be important, so it is useful to also
have its influence metric.
* imatrix: separate metric for attention and ffn importance
* Use stripped tensor name, not src0->name
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* llama-bench: enable having different number of threads for tg and pp
* Add -tgb to usage
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding ability to use THP on Linux
* Use the actual page size4 used for mmap also in munmap
* Add -thp to llama-bench
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Repack a model with the quantize tool
* WIP
* Fixed various issues
As we don't have a way to tell if a repacked quant has been modified,
I had to remove the modification at the expense of a slight decrease
in performance. This affects q8_0_r8, q8_KV_r8, q8_k_r8 on Zen4, and
q4_0_r8 on ARM.
* Create wk_b and wv_b as Q8_0_R8 if the wkv_b type is interleaved
* Fix GCC 13.3 compilation error
* Another one
* Add missing include
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* FlashMLA-2: eliminate intermediate f32 tensors
This works on the CPU. PP performance is ~13% better for 16k tokens
and compute buffer is quite a bit smaller.
* FlashMLA-2: enable fast path only on the CPU for now
I did implement the necessary ops on CUDA, but something is
still wrong there, so for now we only use it when running
CPU-only.
* FlashMLA-2: slightly smaller computer buffer size
* Prepare wk_b when loading DeepSeek models (if wk_b is missing)
* Add some comments
* Fix case where wkv_b is quantized with k- or i-quants.
* Fix CUDA
There is an issue with quantized GEMV on CUDA when the left operand
(the matrix) is not contiguous. So, for now, we also create wv_b
during model loading and use that instead of the 3D view of wkv_b.
* FlashMLA-2: avoid conversions to f32 also on CUDA
* Be able to compute for more than 65535 tokens
On CUDA just a quick hack that allows us to cancatenate tensors
with more than 65535 rows along zroth dimension as needed by
FlashMLA-2. Also needed some care in the perplexity tool to
avoid int overflows when evaluating the computed logits.
* Reduce memory usage for FlashMLA-2
Oh, also fix int overflow in the CUDA concat implementation.
It is funny how the llama.cpp 64-bit police has gone (almost) everywhere
and replaced 32-bit ints with 64-bit ints, needed or not,
but hasn't done it where it is actually needed.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* This gives us ~20% TG speedup for DeepSeek on CUDA
* Slightly better
* Also do it for plain (not fused) mul_mat_id
* Guard against numerical precision issues for MLA on CUDA
* imatrix: wv_b <-> wkv_b
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Custom quantization rules with regular expressions
* Add the --custom-q option to the help
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* A better way to measure the cost of ggml_barrier
* Smart expert selection
* Add ser option to llama-bench
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* This reduces compute buffer size for MLA
* This should accomplish it for standard attention
* Much better
* Better concat for contiguous tensors
If all the op does is to concatenate the second tensor
to the first, why would we want to have a loop?
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
The `-mla` command line option turns into an int from a bool.
mla = 0: use standard attention
mla = 1: use MLA with transposed cache
mla > 1: use MLA without transposed cache
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Give the user the option to override where model weights are stored
* Fix ggml_nbytes() problem and cleanup
For a tensor with zero elements ggml_nbytes() was returning
uint64_t::max, and this was causing graph allocation failure.
* Add timing info to CUDA graph evaluation
* Add more timing info
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Fusing MoE up * unary(gate)
* Fusing MoE up * unary(gate): CUDA
We get ~13% speedup for PP-512 and ~2% for TG-128
for DeepSeek-Lite
* On CUDA also fuse MoE down * (up * unary(gate))
in case the MUL_MAT_ID op for the down experts is the next
op in the graph.
* Command line option to enable fused MoE up*unary(gate)
* Add fmoe option to llama-bench
* Adding forgotten gelu, relu, silu on ARM
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* examples : add new sweep-bench benchmark
* Change documentation to reference ik_llama.cpp
* Made it compile with ik_llama
* Fix JSONL output
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* Adding q8_KV - Basics + AVX2 gemm/gemv
* q8_KV: Better AVX2 gemm
* q8_KV: Better Zen4 gemm
We get 225.7 t/s for L3-8B. In comparison q8_0 without
run-tinme-repacking is at 169 t/s.
* q8_KV: AVX2 gemm/gemv
We get 254 t/s for L3-8B vs 194 t/s for q8_0 without rtr.
* q8_KV: be able to use it for K cache
This required quite a few fixes in ggml and llama.cpp:
* ggml: do not calculate row size as n/block_size*type_size. I had
removed most of it when implementing the quants with per row scale,
bit it was stull lurking in ggml_copy. Not sure if these were the last
remnants of ggmil-style row sizes, or if there are still places left
* llama.cpp: get rid of the the 1d K cache assumption. Create and manage
the K-cache as a 2D tensor so we can have per row meta data as needed
by q8_KV.
Using q8_KV for K-cache results in non-negligible performance gains.
More details to follow, but for DeepSeek-Lite with MLA, we get
18% speedup for PP-8192 compared to q8_0 K-cache.
* q8_KV: be able to use it for K cache in FA
* q8_KV: repack it for K*Q in FA
* q8_KV: slightly faster gemv on Zen4
* q8_KV: slightly faster gemv on Zen4
* q8_KV: ARM_NEON
We get PP-512 = 167 t/s for L3-8B without interleaving!
We do the interleaving on the fly, so I wonder if this
could be done for other quants as well.
* q8_KV: use it in FA on NEON
* q8_KV_r8 - repacked q8_KV
On Zen4 it is slower than q8_k_r8 (292 vs 370 t/s)
This makes no sense whatsoever as the q8_KV_r8 GEMM is
basically the q8_k_r8 GEMM with the unnecessary block stuff
removed (so, one would think that it would be faster).
* q8_KV_r8: don't use nrc_y = 16 on Zen4
This is faster - 350 t/s. Why?
Much better than the 290 t/s we had before, but still slower
than the 370 t/s for q8_k_r8.
* q8_KV: nrc_y = 16 also doesn't pay off in FA
* Minor
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Load all MoE experts during warmup
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* Unify warmup to one token
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* Deepseek MLA Optimizations
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* Make MLA optional
* Remove some unnecessary copies in the MLA attention
* Deepseek MLA Optimizations V2 (#195)
* Avoid allocating MHA KV cache when MLA is turned on
* Added missing gguf-py file
* Added final optimizations
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
* Make sure we do have wk_b and wv_b before enabling MLA
---------
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Use type_k and type_v to set the types of the MLA caches
They were hard-coded at f16.
On my Ryzen-7950X with native bf16 support I get a fairly
significant PP performance boost with bf16 KV-cache:
PP-4096 = 320 t/s up from 292 t/s with fp16 KV-cache.
* Better gemm strategy when nth > nhead
It gives a ~10% PP performance boost for DeepSeek-Lite with 32 threads
(with or without MLA).
Before this commit, when nth > nhead heads were processed
sequentially with all nth threads participating in each
matrix multiplication. Now we ind the gcd of nhead and
nth and split threads into nth/gcd groups, each group
processing nhead/gcd heads.
---------
Co-authored-by: Saood Karim <saood05@gmail.com>
Co-authored-by: Stanisław Szymczyk <sszymczy@gmail.com>
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Rename q4_0_r4 to q4_0_r8 to reflect actual row interleaving
* Rename q8_0_r4 to q8_0_r8 to reflect actual row interleaving
* Rename iq4_xs_r4 to iq4_xs_r8 to reflect actual row interleaving
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq1_m_r4: basics (quantize/dequantize)
* iq1_m_r4: Zen4 gemm
* iq1_m_r4: neon gemm
* iq1_m_r4: switch to q8_0_x4 also on AVX2/Zen4
With the deltas being per group of 8, we cannot make use
of the q8 sums stored in q8_1, so we get a tiny gain by
using q8_0_x4.
* iq1_m_r4: rename mul_mat_iq1_m_r4_q8_1 to mul_mat_iq1_m_r4_q8_0
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq1_s_r4: basics - quantize/dequantize
* iq1_s_r4: gemm/gemv works on AVX2/Zen4
* Don't forget to make sure we have a multiple of 4 rows per thread
* iq1_s_r4: this is better
* iq1_s_r4: fix Zen4 after AVX2 changes
* iq1_s_r4: NEON gemm/gemv
* iq1_s_r4: more bits for shared experts
With this mix we arrive at PPL(512) = 9.4140
for Deepseek-Lite using 1.766 bpw for the repeating layers.
On the Ryzen-7950X we get PP-512 = 494 t/s and
TG-128 = 52 t/s @ 16 threads.
* Forgotten counter increment
* iq1_s_r4: slightly faster AVX2/Zen4 gemm/gemv
* Compiler warnings
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Slightly faster AVX2 implementation for q4_k_r4
* Even better AVX2 implementation for q4_k_r4
We now arrive at PP-512 = 328 t/s for LLaMA-3.1-8B on a
Ryzen-5975WX CPU, up from 291 t/s when I last measured
on 3c5f8722.
With FA and Q8_0 K-cache we get to 339.5 t/s.
* Fix llama-bench labels that I broke with #181
* Faster AVX2 implementation for q5_k_q4
We arrive at 302 t/s for LLaMA-3.1-8B on a Ryzen-5975WX CPU,
up from 273 t/s.
* Use AVX2 implementation of q4_k_r4 and q5_k_r4 also on Zen4
After the changes I made to AVX2, it ends up being slightly faster
compared to what I had for Zen4.
* Minor tweak
* Cleanup
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding gp option to llama-bench
Similar to pg, but it only looks at TG speed with a given
prompt length.
* Make q8_0_r4 work with tensor row sizes that are not a multiple of 128
They still need to be divisible by 32.
* Make q8_0_r4 work with tensor row sizes that are not a multiple of 128
.. on NEON
* Make q8_0_r4 work with tensor row sizes that are not a multiple of 128
.., on AVX2
* Make q4_0_r4 work with tensor row sizes that are not a multiple of 128
.., on AVX2
* Make q4_0_r4 work with tensor row sizes that are not a multiple of 128
... on NEON
* Make q4_0_r4 work with tensor row sizes that are not a multiple of 128
... on Zen4.
Also fix q8_0 K-cache for head sizes that are not multiple of 128.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Fix bug in iqk_mul_mat
I recently added the possibility to have a matrix multiplication
kernel that processes 16 columns in the right matrix per iteration.
This introduced a bug that shows up when batch size is greater
than 16, is not a multiple of 16, and the remainder is not a multiple
of the maximum columns being processed by the regular kernels
(and so, never showed up in my testing using TG-128 and PP-512).
This commit fixes the issue.
* Make sure rows per thread is a multiple of 4 also for MoE when using _r4 quants
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq3_xxs_r4: 1st shot on Zen4
PP-512: 107 t/s -> 137 t/s
TG-128(1 thread): 2.64 t/s -> 3.44 t/s
* iq4_xxs_r4: WIP
* iq4_xxs_r4: 1st shot at AVX2
Note: there is a bug in the AVX2 implementation for nrc_y = 1
for IQ quants with blocks of 32. I have fixed it for now by
using the nrc_y > 1 implementation (which works) also for nrc_y = 1.
* iq3_xxs_r4: NEON
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq5_k_r4: Zen4
Much slower than the others.
* iq5_k_r5: WIP
* Minor
* iq5_k_r4: fix AVX2 nrc_y = 1 case
* iq5_k_r4: better Zen4
But TG is still slower than iq5_k
* iq5_k_r4: slightly better AVX2
* iq5_k_r4: NEON
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Slightly better matrix x vector on Zen4/AVX2 for iq2_k_r4, iq3_k_r4, iq4_k_r4
More importantly: simplify.
* Minor
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Be able to repack tensors at run time
* Repack: also add bf16 as repackable type
* Repack: make sure number of rows is a multiple of the packing
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Not working bf16_r4
* Adding bf16_r8
Small performance gain compared to bf16 - 258 t/s vs 234 t/s.
I guess, this is still sub-obtimal.
* bf16_rx: Very slightly faster by interleaving 16 rows
258 t/s -> 263 t/s
* Rename bf16_r4 to bf16_r16
We are interleaving 16 rows now.
* Cleanup unused stuff
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* q8_k_r8: fastest matrix multiplication known to human kind
We get PP-512(LLaMA-3.1-8B) = 370 t/s on a Ryzen-7950X!
* q8_k_r8: AVX2
I was worried that we don't have enough vector registrers on
AVX2, but it looks like it handles it just fine. We get
PP-512(LLaMA-3.1-8B) = 354 t/s on a Ryzen-5975WX.
Slightly slower than the Zen4 version with double the threads,
but still a huge upgrade compared to Q8_0_R4.
* q8_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 159.2 t/s.
Compare this to the 128 t/s we have fr Q8_0_R4.
* q8_k_r4: go to signed ints
Why?
* On AVX2 _mm256_maddubs_epi16() may overflow, so we need to
stay within the signed int range and use _mm256_sign_epi8.
Not yet tested on the AVX2 comp, vut expect major slowdown.
* It is almost 10% faster on ARM_NEON. Somehow the veorrq_u8()
needed tto convert from unsigned to signed seems to be extremely
slow on the M2-Max
* We only lose ~0.5% in oerformance on Zen4 (there the exclusive
or that we now use to convert fro signed to unsigned seems to be
much faster than on M2-Max)
* Shutup useless compiler warnings
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq4_k_r4: WIP
* iq4_k_r4: Zen4 and hopefully AVX2
On Zen4 we get PP-512(LLaMA-3.1-8B) = 232.6 t/s, up from 182.2 t/s
for iq4_k. Applying the extra shift costs a ~6 performance penalty.
* iq4_k_r4: AVX2
PP-512 = 227.60 t/s. The shifts are really costly.
* iq4_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 108 t/s, up from 58.2 t/s for iq4_k.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* q2_k_r4: Zen4
PP-512(LLaMA-3.1-8B) = 256 t/s
* q3_k_r4: AVX2
* q2_k_r4: AVX2
We get PP-512(LLaMA-3.1-8B) = 287 t/s.
Also cherry-picked the q3_k_r4 AVX2 adaptation that I somehow
forgot to push upstream.
* q2_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 106.2 t/s.
TG-128 is 36.02 t/s, which is ~10% higher than q2_K_S.
* Make sure rows per thread are a multiple of 4
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* q3_k_r4: Zen4 works, but not as good as it should be
238 t/s, so sloghtly slower than q6_k_r4.
* q3_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 106.9 t/s.
This is 1.93X faster than q3_K_S!
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding q6_k_r4
* q6_k_r4: 1st functional AVX2 version
* q6_k_r4: AVX2 and simple Zen4
"Simple" as in processing 4 instead of 8 rows at once.
On Zen4 we get PP-512(LLaMA-3.1-8B) = 238.3 t/s vs
195.2 t/s for Q6_K. TG-128 @ 1 thread is 7.94 t/s
vs 5.38 t/s for Q6_K.
* q6_k_r4: 1st NEON version
PP-512(LLaMA-3.1-8B) = 78 t/s vs 57.6 t/s for q6_K.
TG-128 is slightly lower rthan q6_K for low number of threads,
becomes very slightly better at 8 threads.
* q6_k_r4: slightly faster NEON
PP-512(LLaMA-3.1-8B) = 83.25 t/s
* q6_k_r4: slightly faster Zen4
238.3 t/s -> 243.2 t/s
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Something is still wrong
* Simply don't see what is wrong
* q4_k_r4: finally works on Zen4
I had forgotten to prevent token_embd.weight being quantized
with q4_k_r4!
* q4_k_r4: AVX2
We get PP-512(LLaMA-3.1-8B) = 267 t/s on a Ryzen-5975WX.
This is ~30% better than Q4_K_S.
* q4_k_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 110 t/s.
Not quite as good as q4_0_r4, but still a massive
improvement compared to he 69 t/s for q4_K.
* q4_k_r4: slightly better AVX2
PP-512 goes from 267 t/s to 282 t/s on Ryzen-5975WX
* Minor
* Minor
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding iq2_bn_r4
This Zen4-only implementation achieves PP-512 = 826 t/s (!!!)
for Bitnet-1.58b-3B, up from 620 t/s for iq2_bn.
* Make sure rows per thread are a multiple of the number of interleaved rows
With this I can run iq2_bn_r4 with 32 threads and this increases
PP-512 to 872 t/s.
* iq2_bn_r4: 1st shot at NEON
PP-512 is already faster than iq2_bn (284 t/s vs 246 t/s
for Bitnet-1.58b-3B). TG-128 is ~5% slower.
* iq2_bn_r4: NEON
PP-512 is now 296 t/s. TG-128 is ~20% faster than iq2_bn
for 1 thread, but saturates to about the same 93 t/s at
8 threads.
* iq2_bn_r4: Experimenting on NEON
The matrix x vvector multiplication is erratic.
iq2_bn_r4 is faster at 1, 2, and 4 threads, but
saturates to a lower t/s at 8 threads compared to
iq2_bn. iq2_bn actually manages 99 t/s at 8 threads
and not 93 as I wrore in the last commit. iq2_bn_r4
performance has huge fluctuations at 4 and 8 threads.
* Some cleanup
* iq2_bn_r4: AVX2
As expected, PP is slightly slower as we just don;t have
enough vector registers (690 vs 710 t/s). TG is slightly faster
(18.2 vs 16.7 t/s at 1 thread).
* iq2_bn_r4: use AVX2 implementation on Zen4 for matrix x vector
It is faster - we get 29.6 t/s at 1 thread vs 25.9 t/s for iq2_bn.
* iq2_bn_r4: simdify q8_K16 quantization (AVX2)
PP-512 becomes 834 t/s and TG-128 now saturates to the same
performance as iq2_bn for 4 threads.
* iq2_bn_r4: simdify q8_K16 quantization (NEON)
PP-512 is now 304.7 t/s, and TG-128 @ 8 threads
very slightly outperforms iq2_bn (100.7 t/s vs 99.6 t/s)
* iq2_bn_r4: fix AVX2 after breaking it two commits ago
* iq2_bn_r4: better AVX2
As we don't have enough vector registers on AVX2, it is better
to do two passes per row needing only half of the accumulator
registers that way.
With this, we now beat iq2_bn PP also on AVX2 by a small margin.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding iq4_xs_r4
This is a 1st working version on Zen4.
We get PP-512(LLaMA-3.1-8B) = 226 t/s, so 16% slower
than iq4_nl_x4.
* iq4_xs_r4: WIP
* iq4_xs_r4: Use AVX2 version for matrix x vector on Zen4
* iq4_xs_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 115.6 t/s on M2-Max,
up from 68.2 t/s for iq4_xs!
* DRY
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding q6_0_r4
We get PP-512(LLaMA-3.1-8B) = 257 t/s on a Ryzen-7950X.
* q6_0_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 95 t/s on M2-Max.
In terms of ops, q6_0_r4 is identical to q5_0_r4
except for loading the high bits being
vld1q_u8_x2 instead of vld1q_u8. It is strange that
this can make a 5% difference in performance, especially
considering that this is amortized (re-used) over 8 columns
in the right matrix. Or am I running out of vector registers?
* Fix AVX2
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding q5_0_r4
We get PP-512(LLaMA-3.1-8B) = 256.7 t/s on a Ryzen-7950X.
We even get TG-128 improvement to 11.7 t/s from 11.1 t/s.
* q5_0_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 99.6 t/s on M2-Max,
up from 71.0 t/s for Q5_0. The difference to mainline llama.cpp
is no longer funny: they get 26.5 t/s for Q5_0.
For TG, we are nor able to fully saturate memory bandwidth
and arrive at 22.1 t/s @ 8 threads. Mainline llama.cpp gets
20.6 t/s for Q5_0.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding q8_0_r4
We get PP-512(LLaMA-3.1-8B) = 268 t/s on a Ryzen-7950X compared
to 175.6 t/s for Q8_0.
* q8_0_r4: NEON
We get PP-512(LLaMA-3.1-8B) = 112.6 t/s on M2-Max.
* q8_0_r4: Zen4 matrix-vector specialization
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding iq4_0_r4 - q4_0 repacked
We get PP-512(LLaMA-3.1-8B) = 278 t/s on a Ryzen-7950X CPU,
so ~5-6% faster than iq4_nl_x4.
* q4_0_r4: NEON
Here we get 115.8 t/s, so also ~5% better than iq4_nl_x4.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding iq4_nl_x4
Looks very promising - I get PP-512(LLaMA-3.1-8B) = 230 t/s
on the Ryzen-7950X! This is faster than any other quant and
~40% faster than iq4_nl.
* iq4_nl_x4: getting amazing
This Zen4 variant gets us to PP-512(LLaMA-3.1-8B) = 263 t/s!
* iq4_nl_x4: AVX2
Here we gain only 25% compared to iq4_nl
* iq4_nl_x4: NEON
On M2-Max we get PP-512(LLaMA-3.1-8B) = 109.7 t/s, up from
82.4 t/s for iq4_nl.
* iq4_nl_x4: minor NEON improvement and cleanup
This gets us to 110.3 t/s. In comparison,
IQ4_NL_4_4 in mainline llama.cpp achieves 92.3 t/s.
* iq4_nl_x4: NEON specialization for matrix x vector
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adapting iq2_bn to work without separate scale tensors
Why? It is becoming burdensome to maintain the special Bitnet
conversion in convert_hf_to_gguf.py, so I thnk it is better
to make iq1_bn and iq2_bn just work with the mainline
conversion script (which does not generate scales).
* Adapting iq1_bn to work without separate scale tensors
* Adapting iq2_bn: CUDA dequantize
* Adapting iq2_bn: CUDA works
* Adapting iq1_bn: CUDA works
* Adapting iq1_bn, iq2_bn: NEON
* Adapting iq1_bn, iq2_bn: Metal
Dequantize works, but there is still something wrong
with the dot products.
* WIP
Absoolutely don't see what is wrong with the iq1_bn and iq2_bn
vector dot product kernels.
* Remove iq1_tn and iq2_tn - Part 1
Now that iq1_bn and iq2_bn have per row scales, there is no
reason to also have iq1_tn and iq2_tn.
* Remove iq1_tn and iq2_tn - Part 2
* Bitnet: use the standard llm_build_kv to build self attention
My main motivation was to enable FA. But FA does not work anyway
because head size is 100 for the Botnet ternary models
(and I had forgotten this little detail).
* Revert "Avoid rebuild of GGML graph for each token (#98)"
This reverts commit f2d315b46f7aacc7df4b86bd8acba387b30e11ca.
As far as I can tell, the commit breaks Metal TG.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
To complement the token_embd.weight and output.weight :
attn_v.weight
attn_k.weight.
attn_q_weight
attn_output.weight
attn_qkv.weight
ffn_gate
ffn_down
ffn_up
* iq4_kss: WIP
* iq4_kss: CUDA dequantize works
So we can run perplexity. Sadly, the result does not look good
on the bpw vs quantization error plot.
* iq4_kss: slightly better quantization
* iq4_kss: another small quantization improvement
* iq4_kss: CUDA works
TG-128 performance is very decent with 131 t/s for LLaMA-3.1-8B.
In comparison, we have 123 t/s for q4_0 and 128 t/s for iq4_ks.
I.e., the reduced model size more than offsets the additional
bit fiddling required for iq4_kss.
* iq4_kss: new bit arrangement - CUDA and Zen4 work
Did not lose performance on CUDA. Zen4 is decent, but not great:
PP-512(LLaMA-3.1-8B) = 163 t/s.
TG-128 is of course better than other 4-bit quants due to smaller model size.
We get 14.5 t/s @ 8 threads.
* iq4_kss: ARM_NEON. Predictably very slow
* iq4_kss: Metal
PP is not too bad - just 10% slower than q4_0.
But TG is 30% slower, i.e., predictably bad.
* iq4_kss: somewhat faster Metal dot product
45.75 t/s -> 48.75 t/s.
Still 22% slower than q4_0
* iq4_kss: AVX2
Bad, but better than I expected.
PP-512(LLaMA-3.1-8B) = 167 t/s on the Ryzen-5950X.
I.e., with 32 AVX2 threads we get the performance of
16 Zen4 threads.
* iq4_kss: very slightly faster Metal dot product
48.7 t/s -> 49.3 t/s
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq4_k_xxs: basics
* WIP + adding iq3_kl quantization mix
* iq4_xxs: this looks very viable compared to iq4_xs
At the same 4.25 bpw PPL is always better, for some models
significantly better. I'll rename to iq4_ks and keep it.
* iq4_xxs: CUDA dot product
We get TG-128 = 126 t/s for LLaMA-3.1-8B, compared to 123 t/s for q4_0.
* iq4_xxs: scalar CPU dot product
Also fix the breakage I caused with the dedicated work buffer
quantization portion when the multiplication is not done
via iqk_mul_mat.
* iq4_xxs: Zen4
I noticed that iq4_xs is wrong on Zen4 (and possibly AVX2).
Again the same mistake of packing int32_t back to int16_t,
which overflows occasionally (just occasionally, that's why the
result doesn't look completely wrong, so I didn't notice).
* Fix iq4_xs (Zen4)
* iq4_xxs: AVX2
* iq4_xxs: ARM_NEON
* iq4_xxs: Metal
* iq4_xxs: slightly faster TG on Metal
* iq4_xxs: rename to iq4_ks
After all, tt is a smaller variant of iq4_k.
* iq3_kl: use iq4_ks instead of iq4_k/iq4_xs
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding q6_0 - basics + AVX2/Zen4 working
* Adding q6_0: CUDA dequantize works, but not mmvq
* Adding q6_0: CUDA mmvq works
* Adding q6_0: CUDA cpy, so Q6_0 can be used for KV-cache
* Add q6_0 to CPU flash attention
Disappointing result: for LlaMA-3.2-1B, q6_0 K- and V-cache
gives about the same PPL as q8_0 K-cache and q4_0 V-cache,
while needing the exact same RAM.
I.e., what was the point?
* q6_0: slightly better kv-cache result
Better than q8_0+q4_0, but not as good as q8_0+iq4_nl
* q6_0: works on ARM_NEON
* q6_0: dequantize works on Metal, but not vector dot product
* q6_0: it now works on Metal
Outperforms q5_0 by a significant margin. E.g.
| model | size | params | backend | ngl | threads | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | ------: | ------------: | ---------------: |
| llama 8B Q6_0 | 6.08 GiB | 8.03 B | Metal | 100 | 4 | tg128 | 44.02 ± 0.08 |
| llama 8B Q5_0 | 5.21 GiB | 8.03 B | Metal | 100 | 4 | tg128 | 40.13 ± 0.12 |
| llama 8B Q6_0 | 6.08 GiB | 8.03 B | Metal | 100 | 4 | pp512 | 500.55 ± 0.32 |
| llama 8B Q5_0 | 5.21 GiB | 8.03 B | Metal | 100 | 4 | pp512 | 448.02 ± 0.27 |
* q6_0: can now be used for kv-cache on Metal
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* POC: per row scale
This is a POC how to work around opinionated ggml to
have scales per row rather than per block.
Only implemened for Zen4 and only for iq2_tn.
* POC per row scale: iq2_tn on NEON
* POC per row scale: iq2_tn on Metal
* Per row scale Metal templates
* iq1_tn: shrink to 1.625 bpw (NEON and Metal)
* POC per row scale: CUDA
* POC per row scale: add CUDA TODOs
There are two places in ggml-cuda.cu left where it is assumed
that type_size * n_per_row / block_size is the way to compute
and handle row sizes. This does not affect simple usage,
but will lead to issues when tensors are split between GPUs.
* Per row scales - CUDA
The only place left where there are unnecessary assumptions being made
is in the Flash Attention code. As we are not using any quants that
use per row scales for quantized KV cache, it should be OK for now.
* Update IQ1_TN and IQ2_TN bpw shown to user
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Adding iq1_tn - 1.6875 bpw for TriLM ternary models
* iq1_tn: NEON
* iq1_tn: faster NEON
* iq2_bn: improve performance on NEON
We now get TG-128 = 100 t/s for Bitnet-3B-1.58b!
* iq1_tn: improve AVX2
PP-512 goes to 533 t/s up from 455.
TG-128 @ 2 threads goes to 16.6 t/s up from 14.2.
However, we seem to have a bottleneck somewhere as
TG saturates at 8 threads.
* iq1_tn: improve Zen4
PP-512 goes to 485 t/s up from 352. With FA we get 545 t/s up from 380.
TG-128 @ 1 thread goes to 12.4 t/s up from 10.4.
However, we seem to have a bottleneck somewhere as
TG saturates at 8 threads.
* iq2_bn: improve on Zen4
We now get PP-512 = 614 t/s up from 542 t/s
* iq2_bn: improve AVX2 implementation
We now get PP-512 = 753 t/s up from 680 t/s.
* Remove unnecessary barrier in ggml_compute_forward_mul_mat
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Zen4 Flash Attnetion: WIP bf16
* Zen4 Flash Attnetion: bf16 seems to be working
* Zen4 Flash Attnetion: improving bf16
* Zen4 Flash Attnetion: improving bf16
It is better (slightly faster) to first convert Q
to bf16 before processing each block of q_step rows.
This requires D*q_step*sizeof(bf16) bytes, so at
most 4 kb for the head sizes we support, so we can
just allocate on the stack instead of reserving and
passing a work buffer in ggml.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Softcap: WIP
Fuses scale + tanh + scale as used for softcaping in some
models.
Just CPU for now. ~1.4% for PP-512 on Gemma2-9b, no effect on TG.
Somewhat surprisingly the improvement does not increase as I
go to longer contexts. Gemma2 does softcap on K*Q, which grows
quadratically with context length, so I would have thought
the benefit from fusing scale, tanh, scale would increase.
But no, no luck.
* softcap: CUDA
* softcap: CUDA
~1% speedup for Gemma2-9b
* softcap: Metal and NEON
About 1% speedup.
* Simdified gelu
Gives ~1% speedup for Gemma2-9b prompt processing on AVX512/AVX2.
It looks like the gelu operation is memory bound on my CPU's
after SIMD-ifying it. By not using the 128 kb gelu lookup table
we gain a small advantage.
On the M2-Max the lookup table is slightly faster than the SIMD
version, so left the lookup table for ARM_NEON.
* softcap, tanh: avoid NaNs for large arguments (AVX2, AVX512)
Not that I have encountered this in practice, but just to be sure.
This does it for AVX512 and AVX2, still need a guard for ARM_NEON.
* llama-bench: add ability to turn off warmup runs
So we don't need to wait forever on, e.g., benchmarks involving
long contexts.
* softcap, tanh: avoid NaNs for large arguments (NEON)
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This allows for a better comparison between different models
or different tensors of the same model where the magnitude of
the model weights may differ.
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq2_tn: TriLM specific 2.0625 bpw quantization
Quantize/dequantize/scale dot product.
I get 46 t/s for the TriLM-3.9B with any SIMD!
Finally a compiler doing a decent job auto-vectorizing the
scalar implementation.
* iq2_tn: AVX512
Just reusing the k-quants template gets us to PP-512 = 376 t/s,
TG-128 = 47.6 t/s for TriLM-3.9B.
* iq2_tn: AVX512
With this tweak we get to PP-512 = 431 t/s.
* iq2_tn: AVX512
With this tweak we get TG-128 = 19.58 / 35.18 t/s for 1 / 2 threads.
At 4 threads we saturate at 48.41 t/s, and then performance slowly
degrades with increasing number of threads.
* iq2_tn: AVX2
PP512 = 440 t/s on the Ryzen-5975WX.
We should be able to do better.
* iq2_tn: initial NEON version
* iq2_tn: NEON
For TriLM-3.9B running on the M2-Max we get PP-512 = 193.5 t/s,
TG-128 = 75.5 t/s. This is in line with what we have for
iq2_bn ant 3.3B Bitnet.
* iq2_tn: Metal
For TriLM-3.9B on a 30-core M2-Max we get PP-512 = 890 t/s,
TG-128 = 98.5 t/s.
* iq2_tn: CUDA
For TriLM-3.9B running on RTX-4080 we get PP-512 = 9936 t/s,
TG-128 = 299.2 t/s.
* iq2_tn: AVX2 PP improvement
We now get PP-512 = 490.73 t/s for TriLM-3.9B on the Ryzen-5975WX.
We have PP-512 = 636.61 t/s for Bintnet-3B quantized with iq2_bn.
Bintnet-3B is actually 3.4B, TriLM-3.9B is 3.99B, so we would
expect 3.43/3.99 * 636 = 546 t/s, so it seems we still have something
that is not quite optimal in iq2_tn.
* iq2_tn: small NEON improvement
For TriLM-3.9B we now get PP-512 = 206.6 t/s and TG-128 = 76.4 t/s.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* iq4_k: basics
* quantize/dequantize works
* CUDA dequantize works and one can run PPL calcs. I get
PPL = 6.5258 for LlaMA-3.1-8B, which is 1.77% above fp16.
In comparison, q4_K_S (same size) is 2.88% above fp16.
* TG on CUDA does not work. Johannes has changed the way i-quant dot
products are done, so need to sort out what he had in mind
* iqk_mul_mat is not implemented.
* iq4_k: TG now works on CUDA
* iq4_k: AVX512 implementation
For LLaMA-3.1-8B we get PP-512 = 182.6 t/s, TG-128 = 13.6 t/s,
so almost the same as q4_K_S.
* iq4_k: AVX2 implementation
For LLaMA-3.1-8B we get PP-512 = 203.1 t/s, TG-128 = 12.9 t/s
on the Ryzen-5975X.
* iq4_k: NEON implementation
For LLaMA-3.1-8B we get PP-512 = 60.7 t/s, TG-128 = 25.0 t/s
on the M2-Max. TG is on par with q4_K_S, PP is ~10% slower.
* iq4_k: Metal implementation
For LLaMA-3.1-8B we get PP-512 = 445 t/s, TG-128 = 46.3 t/s
on a 30-core M2-Max GPU. This is to be compared with (currently)
PP-512 = 460 t/s, TG-128 = 51 t/s for q4_K_S.
* iq4_k: scalar dot product
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* Merging mainline - WIP
* Merging mainline - WIP
AVX2 and CUDA appear to work.
CUDA performance seems slightly (~1-2%) lower as it is so often
the case with llama.cpp/ggml after some "improvements" have been made.
* Merging mainline - fix Metal
* Remove check
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
For some models the same tensor is used for token embeddings and
output. This tensor tends to be named token_embedding.weight rather
than output.weight, which prevernts us from collecting imatrix data
for this tensor. With this commit we can tell the name of the
output tensor to the imatrix tool.
* create append_pooling operation; allow to specify attention_type; add last token pooling; update examples
* find result_norm/result_embd tensors properly; update output allocation logic
* only use embd output for pooling_type NONE
* get rid of old causal_attn accessor
* take out attention_type; add in llama_set_embeddings
* bypass logits when doing non-NONE pooling
* cuda sqrt support
* enable cuda in pca
* fix comments in pca
* add test
* add sqrt to ggml_backend_cuda_supports_op
* fix test
* new line
* Use F32 sqrtf instead of F64 sqrt
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* add control-vector-generator
* calc diff
* add comments
* proof-of-concept stdlib implementation
Implements PCA and file writing using mostly standard libraries. The output is recognized as a functional control vector, but outputs gibberish.
* param parsing, refactor, comments
Added basic command-line parameters for outfile and one each positive/negative prompt.
Refactored some messy code in PCA computation and GGUF exporting.
Left a bunch of comments regarding further work needed.
* example template completions
Implements an example template set built from the positive/negative prompts like the control vector Python implementation.
* add multi prompts, multi-thread for PCA
* fix mem error
* add debugs
* fix matrix transpose multiplication
you have got to be kidding me
* preliminary template/multiprompt support
model is running out of context and that ought to be fixed (segfaulting) but other than that it looks goodish
* fix zero output & param parsing, functional templating
fixed a bug where the output file had no tensor data/was all zero
fixed a bug where single hyphen flags were not being correctly parsed
implements creation of templated prompts from input (still need to adapt based on model)
* fix square_diff matmul index range and CRLF->LF line endings
fixed a logic error where square_diff would not multiply all rows
fixed a formatting error where the provided completions.txt had CRLF line endings
* add command-line args for num threads, num completions file lines, always reload model
refactored a few things and did what the commit message says on the tin
* code aestheticization
* fix compiler warnings
* in-series multithreading for prompt embedding?
added commented-out code to attempt to start implementing mutlithreading for embedding in main
* remove unnecessary multithreading
* interim fix memory leak
* translated everything but PCA (I think)
* tentatively translate the rest
* fix ggml errors and make new ones
at least it compiles and runs
* fix cb_eval
* temporary commit while I move dev environments
it finally outputs a functioning control vector - "functioning" in the sense that it can be loaded and it clearly has the right idea, but makes the model incoherent
* update debug statements
* pre-tokenize so we can allocate correct memory to ctx_diffs_wrapped
* update comments
* (wip) refactor
* clean up PCA ggml implementation
* fix shape of v_diff_original
* add n_batch for pca
* working version
* remember to copy back the last_eigenvector
* fix n_completions
* bring back n_completions
* default n_pca_batch to 20
* fix macos build
* add to makefile all targets
* use ggml_format_name
* add readme
* fix .editorconfig
* use ggml_backend_tensor_copy
* attemp to fix compile problem on mac
* fix compile warn
* reuse allocr
* move param parser to common
* better error handling
* clean up a bit
* add print_usage
* shorten help msg
* beautify help msg
* escape prompt by default
* change compile target to llama-cvector-generator
* typo
* disable GPU for PCA
* code style
---------
Co-authored-by: Christian Zhou-Zheng <christianzhouzheng@gmail.com>
* move BLAS to a separate backend
* rename GGML_USE_OPENBLAS to GGML_USE_BLAS
* alloc : reuse same buffer when the same buffer type if used multiple times
* set number of threads automatically for openblas and blis
* sched : print assignments when GGML_SCHED_DEBUG env variable is set
* sched : allow ops with weights on an incompatible buffer type
This will cause the weight to be copied to a backend that supports the
op, which is very costly. The weight should have been stored in a buffer
of a backend that can run the op, but llama.cpp cannot do this
automatically at the moment.
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server : Smart selection of available slot using Longest Common Substring
* add usage
* remove trailing whitespaces
* Use Longest Common Prefix (LCP) instead of LCS
* Rename argument
* avoid to get prompt in infill mode and embedding mode
* remove embedding mode
* refactor format
---------
Co-authored-by: wudexiang <wudexiang@bytedance.com>
* common : gpt_params_parse do not print usage
* common : rework usage print (wip)
* common : valign
* common : rework print_usage
* infill : remove cfg support
* common : reorder args
* server : deduplicate parameters
ggml-ci
* common : add missing header
ggml-ci
* common : remote --random-prompt usages
ggml-ci
* examples : migrate to gpt_params
ggml-ci
* batched-bench : migrate to gpt_params
* retrieval : migrate to gpt_params
* common : change defaults for escape and n_ctx
* common : remove chatml and instruct params
ggml-ci
* common : passkey use gpt_params
* ic
* migrate my eary work
* add the belonging stuff: css,favicon etc
* de prompts
* chore: Update HTML meta tags in index.html file
* add api-key css classes
* some necessary fixes
* Add API key CSS classes and update styling in style.css
* clean the code
* move API to the top, rearrange param sliders. update css
* add tooltips to the parameters with comprehensible explanations
* fix FloatField and BoolField tooltips
* fix grammar field width
* use template literales for promptFormats.js
* update const ModelGenerationInfo
* remove ms per token, since not relevant for most webui users and use cases
* add phi-3 prompt template
* add phi3 to dropdown
* add css class
* update forgotten css theme
* add user message suffix
* fix chatml & add llama3 format
* fix llama3 prompt template
* more prompt format fixes
* add more comon stop tokens
* add missing char
* do not separate with new line or comma
* move prompt style
* add hacky llama2 prompt solution, reduce redundancy in promptFormats.js
* fix toggle state localstorage
* add cmd-r prompt et reduce redundancy
* set default prompt to empty
* move files, clean code
* fix css path
* add a button to the new ui
* move new ui to "/public" due to otherwise problematic CORS behaviour
* include new ui in cpp
* fix wrong link to old ui
* renaming to ensure consistency
* fix typos "prompt-format" -> "prompt-formats"
* use correct indent
* add new ui files to makefile
* fix typo
* SimpleChat:DU:BringIn local helper js modules using importmap
Use it to bring in a simple trim garbage at end logic, which is
used to trim received response.
Also given that importmap assumes esm / standard js modules, so
also global variables arent implicitly available outside the
modules. So add it has a member of document for now
* SimpleChat:DU: Add trim garbage at end in loop helper
* SimpleChat:DU:TrimGarbage if unable try skip char and retry
* SimpleChat:DU: Try trim using histogram based info
TODO: May have to add max number of uniq chars in histogram at
end of learning phase.
* SimpleChat:DU: Switch trim garbage hist based to maxUniq simple
Instead of blindly building histogram for specified substring
length, and then checking if any new char within specified min
garbage length limit, NOW exit learn state when specified maxUniq
chars are found. Inturn there should be no new chars with in
the specified min garbage length required limit.
TODO: Need to track char classes like alphabets, numerals and
special/other chars.
* SimpleChat:DU: Bring in maxType to the mix along with maxUniq
Allow for more uniq chars, but then ensure that a given type of
char ie numerals or alphabets or other types dont cross the
specified maxType limit. This allows intermixed text garbage
to be identified and trimmed.
* SimpleChat:DU: Cleanup debug log messages
* SimpleChat:UI: Move html ui base helpers into its own module
* SimpleChat:DU:Avoid setting frequence/Presence penalty
Some models like llama3 found to try to be over intelligent by
repeating garbage still, but by tweaking the garbage a bit so that
it is not exactly same. So avoid setting these penalties and let
the model's default behaviour work out, as is.
Also the simple minded histogram based garbage trimming from end,
works to an extent, when the garbage is more predictable and
repeatative.
* SimpleChat:UI: Add and use a para-create-append helper
Also update the config params dump to indicate that now one needs
to use document to get hold of gMe global object, this is bcas of
moving to module type js.
Also add ui.mjs to importmap
* SimpleChat:UI: Helper to create bool button and use it wrt settings
* SimpleChat:UI: Add Select helper and use it wrt ChatHistoryInCtxt
* SimpleChat:UI:Select: dict-name-value, value wrt default, change
Take a dict/object of name-value pairs instead of just names.
Inturn specify the actual value wrt default, rather than the
string representing that value.
Trap the needed change event rather than click wrt select.
* SimpleChat:UI: Add Div wrapped label+element helpers
Move settings related elements to use the new div wrapped ones.
* SimpleChat:UI:Add settings button and bring in settings ui
* SimpleChat:UI:Settings make boolean button text show meaning
* SimpleChat: Update a bit wrt readme and notes in du
* SimpleChat: GarbageTrim enable/disable, show trimmed part ifany
* SimpleChat: highlight trim, garbage trimming bitmore aggressive
Make it easy for end user to identified the trimmed text.
Make garbage trimming logic, consider a longer repeat garbage
substring.
* SimpleChat: Cleanup a bit wrt Api end point related flow
Consolidate many of the Api end point related basic meta data into
ApiEP class.
Remove the hardcoded ApiEP/Mode settings from html+js, instead use
the generic select helper logic, inturn in the settings block.
Move helper to generate the appropriate request json string based
on ApiEP into SimpleChat class itself.
* SimpleChat:Move extracting assistant response to SimpleChat class
so also the trimming of garbage.
* SimpleChat:DU: Bring in both trim garbage logics to try trim
* SimpleChat: Cleanup readme a bit, add one more chathistory length
* SimpleChat:Stream:Initial handshake skeleton
Parse the got stream responses and try extract the data from it.
It allows for a part read to get a single data line or multiple
data line. Inturn extract the json body and inturn the delta
content/message in it.
* SimpleChat: Move handling oneshot mode server response
Move handling of the oneshot mode server response into SimpleChat.
Also add plumbing for moving multipart server response into same.
* SimpleChat: Move multi part server response handling in
* SimpleChat: Add MultiPart Response handling, common trimming
Add logic to call into multipart/stream server response handling.
Move trimming of garbage at the end into the common handle_response
helper.
Add new global flag to control between oneshot and multipart/stream
mode of fetching response. Allow same to be controlled by user.
If in multipart/stream mode, send the stream flag to the server.
* SimpleChat: show streamed generative text as it becomes available
Now that the extracting of streamed generated text is implemented,
add logic to show the same on the screen.
* SimpleChat:DU: Add NewLines helper class
To work with an array of new lines. Allow adding, appending,
shifting, ...
* SimpleChat:DU: Make NewLines shift more robust and flexible
* SimpleChat:HandleResponseMultiPart using NewLines helper
Make handle_response_multipart logic better and cleaner. Now it
allows for working with the situation, where the delta data line
got from server in stream mode, could be split up when recving,
but still the logic will handle it appropriately.
ALERT: Rather except (for now) for last data line wrt a request's
response.
* SimpleChat: Disable console debug by default by making it dummy
Parallely save a reference to the original func.
* SimpleChat:MultiPart/Stream flow cleanup
Dont try utf8-decode and newlines-add_append if no data to work on.
If there is no more data to get (ie done is set), then let NewLines
instance return line without newline at end, So that we dont miss
out on any last-data-line without newline kind of scenario.
Pass stream flag wrt utf-8 decode, so that if any multi-byte char
is only partly present in the passed buffer, it can be accounted
for along with subsequent buffer. At sametime, bcas of utf-8's
characteristics there shouldnt be any unaccounted bytes at end,
for valid block of utf8 data split across chunks, so not bothering
calling with stream set to false at end. LATER: Look at TextDecoder's
implementation, for any over intelligence, it may be doing..
If needed, one can use done flag to account wrt both cases.
* SimpleChat: Move baseUrl to Me and inturn gMe
This should allow easy updating of the base url at runtime by the
end user.
* SimpleChat:UI: Add input element helper
* SimpleChat: Add support for changing the base url
This ensures that if the user is running the server with a
different port or wants to try connect to server on a different
machine, then this can be used.
* SimpleChat: Move request headers into Me and gMe
Inturn allow Authorization to be sent, if not empty.
* SimpleChat: Rather need to use append to insert headers
* SimpleChat: Allow Authorization header to be set by end user
* SimpleChat:UI+: Return div and element wrt creatediv helpers
use it to set placeholder wrt Authorization header.
Also fix copy-paste oversight.
* SimpleChat: readme wrt authorization, maybe minimal openai testing
* SimpleChat: model request field for openai/equivalent compat
May help testing with openai/equivalent web services, if they
require this field.
* SimpleChat: readme stream-utf-8 trim-english deps, exception2error
* Readme: Add a entry for simplechat in the http server section
* SimpleChat:WIP:Collate internally, Stream mode Trap exceptions
This can help ensure that data fetched till that point, can be
made use of, rather than losing it.
On some platforms, the time taken wrt generating a long response,
may lead to the network connection being broken when it enters
some user-no-interaction related power saving mode.
* SimpleChat:theResp-origMsg: Undo a prev change to fix non trim
When the response handling was moved into SimpleChat, I had changed
a flow bit unnecessarily and carelessly, which resulted in the non
trim flow, missing out on retaining the ai assistant response.
This has been fixed now.
* SimpleChat: Save message internally in handle_response itself
This ensures that throwing the caught exception again for higher
up logic, doesnt lose the response collated till that time.
Go through theResp.assistant in catch block, just to keep simple
consistency wrt backtracing just in case.
Update the readme file.
* SimpleChat:Cleanup: Add spacing wrt shown req-options
* SimpleChat:UI: CreateDiv Divs map to GridX2 class
This allows the settings ui to be cleaner structured.
* SimpleChat: Show Non SettingsUI config field by default
* SimpleChat: Allow for multiline system prompt
Convert SystemPrompt into a textarea with 2 rows. Reduce
user-input-textarea to 2 rows from 3, so that overall
vertical space usage remains same.
Shorten usage messages a bit, cleanup to sync with settings ui.
* SimpleChat: Add basic skeleton for saving and loading chat
Inturn when ever a chat message (system/user/model) is added,
the chat will be saved into browser's localStorage.
* SimpleChat:ODS: Add a prefix to chatid wrt ondiskstorage key
* SimpleChat:ODS:WIP:TMP: Add UI to load previously saved chat
This is a temporary flow
* SimpleChat:ODS:Move restore/load saved chat btn setup to Me
This also allows being able to set the common system prompt
ui element to loaded chat's system prompt.
* SimpleChat:Readme updated wrt save and restore chat session info
* SimpleChat:Show chat session restore button, only if saved session
* SimpleChat: AutoCreate ChatRequestOptions settings to an extent
* SimpleChat: Update main README wrt usage with server
* SimpleChat: A placeholder system prompt, Use usage msg in code
Just have a alert msg wrt needing javascript enabled in html. And
have usage message from js file. Update the usage message a bit.
So also enable switch session wrt setup_ui call.
Add a possible system prompt as a placeholder for the system-input.
* SimpleChat:CompletionMode: Allow control of Role: prefix
* SimpleChat:Completion: Avoid Role: prefix; Newline only in between
In completion mode
* avoid inserting Role: prefix before each role's message
* avoid inserting newline at the begin and end of the prompt
message. However if there are multiple role messages, then
insert newline when going from one role's message to the
next role's message.
* SimpleChat:CompletionMode: Update readme/usage, trim textarea newline
Readme update wrt completion mode behavior.
Usage help updated wrt completion mode behavior.
When changing from input to textarea elment wrt user input, the last
newline at the end of the user input wrt textarea, was forgotten to be
filtered, this is fixed now. However if user wants to have a explicit
newline they can using shift+enter to insert a newline, that wont be
removed. The extra newline removal logic uses substring and keyup to
keep things simple and avoid some previously noted bugs wrt other
events in the key path as well as IME composition etal.
* SimpleChat:SC: Ensure proper clearing/reseting
previous logic would have cleared/reset the xchat, without doing
the same wrt iLastSys, thus leading to it pointing to a now non
existent role-content entry.
So if a user set a system prompt and used completion mode, it would
have done the half stupid clear, after the model response was got.
Inturn when user tries to send a new completion query, it would
inturn lead to handle_user_submit trying to add/update system prompt
if any, which will fail, bcas iLastSys will be still pointing to a
non existant entry.
This is fixed now, by having a proper clear helper wrt SC class.
* SimpleChat: Update usage note and readme a bit
* SimpleChat:Completion: clear any prev chat history at begining
Previously any chat history including model response to a completion
query would have got cleared, after showing the same to the user,
at the end of handle_user_submit, rather than at the begining.
This gave the flexibility that user could switch from chat mode
to completion mode and have the chat history till then sent to
the ai model, as part of the completion query. However this flow
also had the issue that, if user switches between different chat
sessions, after getting a completion response, they can no longer
see the completion query and its response that they had just got.
The new flow changes the clearing of chat history wrt completion
mode to the begining of handle_user_submit, so that user doesnt
lose the last completion mode query and response, till a new
completion mode query is sent to the model, even if they were to
switch between the chat sessions. At the same time the loss of
flexibility wrt converting previous chat history into being part
of the completion query implicitly doesnt matter, because now
the end user can enter multiline queries.
* SimpleChat:Try read json early, if available
For later
the server flow doesnt seem to be sending back data early, atleast
for the request (inc options) that is currently sent.
if able to read json data early on in future, as and when ai model
is generating data, then this helper needs to indirectly update
the chat div with the recieved data, without waiting for the
overall data to be available.
* SimpleChat: Rename the half asleep mis-spelled global var
* SimpleChat: Common chat request options from a global object
* SimpleChat: Update title, usage and readme a bit
Keep the title simple so that print file name doesnt have chars
that need to be removed.
Update readme wrt some of the new helpers and options.
Change Usage list to a list of lists, add few items and style it
to reduce the margin wrt lists.
* SimpleChat:ChatRequestOptions: max_tokens
As some times based on the query from the user, the ai model may get
into a run away kind of generation with repeatations etal, so adding
max_tokens to try and limit this run away behaviour, if possible.
* SimpleChat: Reduce max_tokens to be small but still sufficient
* SimpleChat: Consolidate global vars into gMe, Display to user
This allows the end user to see the settings used by the logic,
as well as allows users to change/update the settings if they
want to by using devel-tools/console
* SimpleChat:SlidingWindow: iRecentUserMsgCnt to limit context load
This is disabled by default. However if enabled, then in addition
to latest system message, only the last N user messages, after the
latest system message and its reponses from the ai model will be sent
to the ai-model, when querying for a new response.
This specified N also includes the latest user query.
* SimpleChat: placeholder based usage hint for user-in textarea
* SimpleChat: Try make user experience better, if possible
Reduce chat history context sent to the server/ai-model to be
just the system-prompt, prev-user-request-and-ai-response and
cur-user-request, instead of the previous full chat history.
This way if there is any response with garbage/repeatation, it
doesnt mess with things beyond the next question, in some ways.
Increase max_tokens to 1024, so that a relatively large previous
reponse doesnt eat up the space available wrt next query-response.
However dont forget that the server when started should also
be started with a model context size of 1k or more, to be on
safe side.
Add frequency and presence penalty fields set to 1.2 to the set
of fields sent to server along with the user query. So that
the model is partly set to try avoid repeating text in its
response.
* SimpleChat:Add n_predict (equiv max_tokens) for llamacpp server
The /completions endpoint of examples/server doesnt take max_tokens,
instead it takes the internal n_predict, for now add the same on
the client side, maybe later add max_tokens to /completions endpoint
handling.
* SimpleChat: Note about trying to keep things simple yet flexible
* main : don't print special tokens with --grammar
The CLI interface was recently changed to print special control tokens
like the </s> stop message one. This token shouldn't be printed if the
grammar flag was passed, unless the grammar specifies it, because that
breaks shell-scriptability.
* main: use seperate stream for control characters
* main: use dprintf and add --ctrl-token-no-out and --ctrl-token-fd-out
* main: dprintf isn't part of the IEEE POSIX standard. Just use write().
* main: remove --ctrl-token-fd-out in favor for fcntl() based detection
* common.cpp: accidentally removed --interactive-first
* main: only merge stdout and control token if not in conversation or grammar mode
* main: rejig control token descriptor handling
* main: must check pipe status on very top of program
* main: renamed --no-special from --ctrl-token-no-out and other refactoring
* main: refactor ctrl_token_no_out --> no_special
* llama: rename llama_token_is_control_token() to llama_token_is_control()
* main: remove special token file descriptor feature (#5)
---------
Co-authored-by: Brian <mofosyne@gmail.com>
* Make tokenizer.cpp CLI tool nicer.
Before this commit, tokenize was a simple CLI tool like this:
tokenize MODEL_FILENAME PROMPT [--ids]
This simple tool loads the model, takes the prompt, and shows the tokens
llama.cpp is interpreting.
This changeset makes the tokenize more sophisticated, and more useful
for debugging and troubleshooting:
tokenize [-m, --model MODEL_FILENAME]
[--ids]
[--stdin]
[--prompt]
[-f, --file]
[--no-bos]
[--log-disable]
It also behaves nicer on Windows now, interpreting and rendering Unicode
from command line arguments and pipes no matter what code page the user
has set on their terminal.
* style fix: strlen(str) == 0 --> *str == 0
* Simplify tokenize.cpp; by getting rid of handling positional style arguments.
It must now be invoked with long --model, --prompt etc. arguments only.
Shortens the code.
* tokenize.cpp: iostream header no longer required
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: brian khuu <mofosyne@gmail.com>
* SimpleChat: Add a skeletal html page
Contains a div placeholder for showing chat messages till now
a text-input for allowing user to enter next chat message/query
to the model.
a submit button to allow sending of the user entered message and
chat till now to the model.
* SimpleChat: A js skeleton with SimpleChat class
Allows maintaining an array of chat message.
Allows adding chat message (from any of the roles be it system,
user, assistant, ...)
Allows showing chat messages till now, in a given div element.
* SimpleChat: request_json, globals, startme
* SimpleChatJS: Roles Class, submitClick
Define Role class with static members corresponding to the roles.
Update startme to
* Get hold of the ui elements.
* Attach a click handler to submit button, which adds the user input
to xchats array and shows the chat messages till now in chat div
element.
Trap DOMContentLoaded to trigger startme
* SimpleChat:HTML: Bring in the js file
* SimpleChat: Rather value wrt input text element
* SimpleChat: Also add completions related prompt
* SimpleChat: Use common helper logic wrt json data
* SimpleChat: Move handling of submit request into its own func
* SimpleChat: Try handshake with llm over its web service endpoint
* SimpleChat:JS: Extract model response and show to user
* SimpleChat:JS: Messages/Prompt, indicate working to end user
* SimpleChat: Try keep input element in view
* SimpleChat: Diff user/assistant msgs, Make input wider
Also show a default message to user
Also add some metas
* SimpleChat: Move into its own sub directory to avoid confusion
* SimpleChat:sh: Add simple shell script to run python3 http.server
So one needs to run the llm server locally
then run this script and access it using a local browser
* SimpleChat:JS: Try trap enter key press wrt input text field
So user can either press submit button or press enter key
* SimpleChat: Allow user to select chat or completion mode
* SimpleChat: Dont submit if already submitted and waiting
Also make chat the default selection wrt mode
* SimpleChat:JS: Handle difference in response
Try read the assistance response from appropriate field in the
response got.
Also examples/server seems to return the response in a slightly
different field, so try account for that also.
* SimpleChat:JS: Force completion mode be single message by default
* SimpleChat: Add a simple readme file
* SimpleChat:HTML: Cleanup/structure UI a bit, Add input for system
* SimpleChat:Allow system prompt to be set, if provided before user
* SimpleChat: Ignore empty user input, without trimming
* SimpleChat:Alert user if they provide sysprompt late or change it
* SimpleChat: Move handling systemprompt into its own func
* SimpleChat:HTML: Add a style for system role message
* SimpleChat: Update the readme file
* SimpleChat:CSS: Move style info into its own css file
To keep it simple, clean and seperate so that things are not
unnecessarily cluttered.
* SimpleChat:CSS: Allow for chat div to be scrollable
* SimpleChat:JS: Try ensure the last entry in chat is visible
Needed because now only the chat div is scrollable and not the full
page.
In last commit the chat div size was fixed to 75% vertical height,
so the full page no longer scrolls, so the old bring user-input
element to view wont work, instead now the last element in the
chat div should be brought into view.
* SimpleChat:JS: bottom of element visible, Set focus to user input
As the generated text could be multiple lines and occupy more space
that the full scrollable div's vertical space, make the bottom of
the last element (which can be such a generated text) in the div
visible by scrolling.
Ensure that the user input box has focus
* SimpleChat: Update notes a bit. Try keep browser happy
Avoid browser quirk mode with DOCTYPE.
Help with accessibility a bit by specifying the language explicitly.
Specify the char encoding explicitly, inturn utf-8 is a safe bet,
even with intermixing of languages if reqd in future.
Add a cache-control http-equiv meta tag, which in all probability
will be ignored.
Defer js loading and execution, just for fun and future, not that
critical here as it stands now.
* SimpleChat:HTML:Group user input+btn together; Note about multichat
* SimpleChat:JS: Allow for changing system prompt anytime for future
* SimpleChat:Readme: Note about handle_systemprompt begin/anytime
* SimpleChat:HTML: Add viewport meta for better mobile friendliness
Without this the page content may look too small.
* SimpleChat:HtmlCss: Cleanup UI flow
set margin wrt vmin rather than vw or vh so portrait/landscape ok.
Use flex and flex-grow to put things on the same line as well as
distribute available space as needed. Given two main elements/line
so it remains simple.
In each line have one element with grows and one sits with a basic
comfortably fixed size.
* SimpleChat: textarea for multiline user chat, inturn shift+enter 4 enter
* SimpleChat: Make vertical layout better responsive (flex based)
Also needed to make things cleaner and properly usable whether
landscape or portrait, after changing to multiline textarea rather
than single line user input.
Avoid hardcoding the chat-till-now display area height, instead
make it a flex-growable within a flex column of ui elements within
a fixed vertical area.
* SimpleChat: Rename simplechat.html to index.html, update readme
Instead of providing a seperate shell script, update the readme wrt
how to run/use this web front end.
* SimpleChat: Screen fixed view and scrolling, Printing full
* SimpleChat:JS:CI: Avoid space at end of jsdoc param line
* SimpleChat:JS: MultiChat initial skeleton
Will help maintain multiple independent chats in future
* SimpleChat:JS: Move system prompt begin/anytime into SimpleChat
* SimpleChat:JS:Keep MultiChatUI simple for now
Worry about different chats with different servers for later.
* SimpleChat:JS: Move handle submit into MultiChat, build on same
Create an instance of MultiChatUI and inturn a instance of chat
session, which is what the UI will inturn work on.
* SimpleChat:JS: Move to dictionary of SimpleChat, instead of array
* SimpleChat: Move ui elements into MultiChatUI, Update el IDs
Move ui elements into MultiChatUI, so that current handleUserSubmit
doesnt need to take the element arguments. Also in future, when
user is allowed to switch between different chat sessions, the
UI can be updated as needed by using the elements in UI already
known to MultiChatUI instance.
Rename the element ids' so that they follow a common convention,
as well as one can identify what the element represents in a more
consistant manner.
* SimpleChat:MCUI:Show available chat sessions, try switch btw them
Previous commits brought in / consolidated existing logic into
MultiChatUI class.
Now start adding logic towards multichat support
* show buttons indicating available chat sessions
* on sessin button click, try switch to that session
* SimpleChat:MCUI: Store and use current chat session id
Also
allow to switch chat session optionally, wrt some of the related
helpers.
setup for two chat sessions by default.
* SimpleChat:MCUI: Delay enabling user-input to avoid race
Re-enable user-input, only after response to a user query has been
updated to the chat-div. This ensures that if user tries to switch
chat session, it wont be allowed till chat-request-response flow is
done.
* SimpleChat: Take care of system prompt
Helper to get the latest system prompt and inturn use same to
set the system prompt ui, when switching.
Ensure that system prompt is set if and when enter key is pressed.
* SimpleChat:GetSystemLatest, fix a oversight.
* SimpleChat:MCUI: Allow selected chat-session btn to be highlighted
Also have a general helper for setting class of children.
* SimpleChat:Cleanup corners
Show system prompt in chat space, when it is set by pressing enter,
as a feedback to user.
Alert user, if they try to switch chat session in the middle of
waiting for a response from the ai model.
* SimpleChat:MCUI: Ensure req-resp failure doesnt lock up things
* SimpleChat:MCUI: Support for new chat sessions
Also a general create button helper.
* SimpleChat:MCUI: CreateSessionBtn helper, use wrt NewChat
Also fix a oversight wrt using stale data wrt the list of chat
sessions.
* SimpleChat:MCUI: NewChat btn first before existing chat sessions
* SimpleChat:MCUI:CornerCases:Skip new chat, show only if current
Skip NewChat if user cancels or if one waiting for response from
the ai model.
Dont show a chat with newly got ai model response, if current chat
session has changed, some how. Chat session shouldnt be allowed to
change, if there is a pending response, but still as a additional
sanity check.
* SimpleChat: Update readme, title, show usage if no chat to show
* SimpleChat: Cleanup the log/dialog messages a bit
* phi3 : duplicate rope factors in each layer
phi3 : set phi-3 model type as 14B
model loader : simplify the process for duplicating model tensors
llama-bench : remove default pg test
* replace bool parameters in llama_model_loader with named flags
* add phi3 128k support in convert-hf-to-gguf
* add phi3 128k support in cuda
* address build warnings on llama.cpp
* adjust index value in cuda long rope freq factors
* add long rope support in ggml cpu backend
* make freq factors only depend on ctx size
* remove unused rope scaling type 'su' frin gguf converter
* fix flint warnings on convert-hf-to-gguf.py
* set to the short freq factor when context size is small than trained context size
* add one line of comments
* metal : support rope freq_factors
* ggml : update ggml_rope_ext API to support freq. factors
* backends : add dev messages to support rope freq. factors
* minor : style
* tests : update to use new rope API
* backends : fix pragma semicolons
* minor : cleanup
* llama : move rope factors from KV header to tensors
* llama : remove tmp assert
* cuda : fix compile warning
* convert : read/write n_head_kv
* llama : fix uninitialized tensors
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* examples: cache hf model when --model not provided
* examples: cache hf model when --model not provided
* examples: cache hf model when --model not provided
* examples: cache hf model when --model not provided
* examples: cache hf model when --model not provided
* Update brute force test: add_special
* Update brute force test: default values for add_bos_token and add_eos_token
* Enable rtrim when pre-inserting BOS
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Revert "server : fix test regexes"
* Update brute force test: special tokens
* Fix added tokens
- Try to read 'added_tokens.json'.
- Try to read 'tokenizer_config.json'.
- Try to read 'tokenizer.json'.
* Fix special tokens rtrim
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* server : fix test regexes
* android : use "ci-android" branch for CI
* ggml : disable SIMD exp and silu for 32-bit ARM
ggml-ci
* android : do not fetch, use add_subdirectory instead
* cmake : provide binary dir
- Change '--embedding' to '--embeddings' in the README
- Update the description to match the latest --help output
- Added a caution about defining physical batch size
* [server] Cleanup a memory leak on exit
There are a couple memory leaks on exit of the server. This hides others.
After cleaning this up, you can see leaks on slots. But that is another
patch to be sent after this.
* make tab into spaces
* feat: first things to do
* feat: create tensors for Jina architecture
* fix: use other tensors
* feat: embedding gets results
* fix: fix usage of ALIBI
* fix: clean prints
* fix: do some cleanup unused vars
* fix: revert changes to Makefile and CMakeLists
* fix: revert some changes
* fix: fix small detail
* fix: fix convert formatting
* fix: fix linting and editor
* feat: set proper vocab settings
* fix: JinaBertForMaskedLM registration
* feat: support q_normalization and k_normalization in Jina arch
* feat: handle gpt2 tokenizer with Jina architecture
* feat: example comments in embedding
* feat: rename Jina Bert to Jina Bert V2
* fix: add some changes as per review
* feat: proper KQ_pos for Jina embeddings
* feat: add capacity to load models ES and DE for Spanish
* llama : fix pre-tokenizers
* ggml : full ALiBi support
* ggml : update ggml_soft_max_ext() CUDA, SYCL
* ggml : ggml_flash_attn_ext() support ALiBi (CPU)
* ggml : ggml_flash_attn_ext() support ALiBi (Metal)
* ggml : fix warning
* ggml : ggml_flash_attn_ext() support ALiBi (CUDA)
ggml-ci
* minor : clean-up
* embedding : add warning about missing SEP
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
The llama.cpp grammar parser had a bug where forgetting to add a closing
quotation mark to strings would cause parsing to crash. Anyone running a
server on a public endpoint is advised to upgrade. To reproduce this bug
./llamafile -m foo.gguf -p bar --grammar 'root::="'
Credit for discovering and reporting this issue goes to Eclypsium
Security Researcher Richard Johnson <Richard.johnson@eclypsium.com>.
* Revert "Revert "llava : add support for moondream vision language model (#6899)""
This reverts commit 9da243b36ac0b9d609adfaaa4c8f1cc8c592f737.
* Fix num_positions and embeddings initialization
* convert-hf : begin refactoring write_tensor
* convert : upgrade to sentencepiece v0.2.0
* convert-hf : remove unused n_dims in extra_*_tensors
* convert-hf : simplify MoE weights stacking
* convert-hf : flake8 linter doesn't like semicolons
* convert-hf : allow unusual model part names
For example, loading `model-00001-of-00001.safetensors` now works.
* convert-hf : fix stacking MoE expert tensors
`torch.stack` and `torch.cat` don't do the same thing.
* convert-hf : fix Mamba conversion
Tested to work even with a SentencePiece-based tokenizer.
* convert : use a string for the SentencePiece tokenizer path
* convert-hf : display tensor shape
* convert-hf : convert norms to f32 by default
* convert-hf : sort model part names
`os.listdir` is said to list files in arbitrary order.
Sorting the file names should let "model-00009-of-00042.safetensors"
be loaded before "model-00010-of-00042.safetensors".
* convert-hf : use an ABC for Model again
It seems Protocol can't be used as a statically type-checked ABC,
because its subclasses also can't be instantiated. (why did it seem to work?)
At least there's still a way to throw an error when forgetting to define
the `model_arch` property of any registered Model subclasses.
* convert-hf : use a plain class for Model, and forbid direct instantiation
There are no abstract methods used anyway,
so using ABC isn't really necessary.
* convert-hf : more consistent formatting of cmdline args
* convert-hf : align the message logged for converted tensors
* convert-hf : fix Refact conversion
* convert-hf : save memory with lazy evaluation
* convert-hf : flake8 doesn't like lowercase L as a variable name
* convert-hf : remove einops requirement for InternLM2
* convert-hf : faster model parts loading
Instead of pre-loading them all into a dict, iterate on the tensors
in the model parts progressively as needed in Model.write_tensors
Conversion for some architectures relies on checking for the presence
of specific tensor names, so for multi-part models, the weight map is read
from the relevant json file to quickly get these names up-front.
* convert-hf : minor changes for consistency
* gguf-py : add tqdm as a dependency
It's small, and used for a progress bar
in GGUFWriter.write_tensors_to_file
* Added themes support with two sample themes and a favicon.
* Newline
* Newline
* Newline
* Trailing whitespace
* Increased opacity for contrast
* Increase opacity.
Check actions cancelled for some other priority job and I can't seem to manually re-run them, so MOAR OPACITY
* Opacity action trigger.
Trying to re-trigger the cancelled action.
* One more opacity adjustment
This Actions pipeline is failing for random issues.
* Delete examples/server/themes/buttons_top/completion.js
This will be served from the static string built-in to server.
* Delete examples/server/themes/buttons_top/index.js
This will be served from the static string built-in to server.
* Delete examples/server/themes/wild/completion.js
This will be served from the static string built-in to server.
* Delete examples/server/themes/buttons_top/json-schema-to-grammar.mjs
This will be served from the static string built-in to server.
* Delete examples/server/themes/wild/index.js
This will be served from the static string built-in to server.
* Delete examples/server/themes/wild/json-schema-to-grammar.mjs
This will be served from the static string built-in to server.
* Replaced underscore.
* Introduce bfloat16 support
Many models on Hugging Face (e.g. Mistral, TinyLLaMA) use bfloat16 as
their canonical floating point format.
┌sign
│
│ ┌exponent
│ │
│ │ ┌mantissa
│ │ │
│┌──┴───┐┌─┴───┐
0b0000000000000000 brain16
This encoding has the same number of exponent bits as float32. That
makes conversion relatively straightforward, even in the absence of
hardware support. For example, converting brain16 to binary32 means
simply shifting 16 bits to the left.
┌sign
│
│ ┌exponent
│ │
│ │ ┌mantissa
│ │ │
│┌──┴───┐┌─┴───────────────────┐
0b00000000000000000000000000000000 IEEE binary32
The issue is that converting bf16 to fp16 can result in information
loss. Only 13% of bf16 numbers can be precisely represented in fp16
which in practice ends up being 99.71% of Mistral 7b v0.2's weights
however there is currently no way other than fp32 to get the others
┌sign
│
│ ┌exponent
│ │
│ │ ┌mantissa
│ │ │
│┌─┴─┐┌─┴──────┐
0b0000000000000000 IEEE binary16
This change fixes that, by adding a bf16 data type to GGML. Support
for CPU inference has been implemented along with optimizations for
the AVX2, AVX512, and AVX512BF16 ISAs. Perplexity on Mistral 7b 0.2
improves somewhere around -0.0024 to -0.0046 compared to using fp16
* Remove GGML code that's not needed
* Minimize the GGML API surface area for BF16
* Remove bf16 luts
* Make the GGML header look nicer
* Fix documentation
* Apply ggerganov's fixes for test-backend-ops
* Add BF16 code for new ggml_validate_row_data() function
* Fixed save_imatrix to match old behaviour for MoE
This fix is simple and clear, but unnecessarily doubles the memory overhead..
* Fixed missing idx variable
* Unconditionally increment ncall
Co-authored-by: slaren <slarengh@gmail.com>
* Fixed 2 bugs in save_imatrix()
- Fixed segfault bug because the counts vector needed to be created.
- Fixed pre-existing bug didn't actually add to the counts for "--combine" option.
* ncall needs summing too
* Trailing whitespace
---------
Co-authored-by: slaren <slarengh@gmail.com>
* Update log text (EOS to EOG)
The log text "found EOS" is no longer always correct, here, because there is now an is-EOG check that also returns true for EOT.
* Improve log msg. further by using "an" instead of "some".
As suggested, to avoid misunderstanding (no multiple EOG tokens found, just one).
This will reproduce the issue in llama13b
{
'prompt': 'Q: hello world \nA: ',
'stop': ['\n'],
'temperature': 0.0,
'n_predict': 10,
'cache_prompt': True,
'n_probs': 10
}
* ggml : add ggml_flash_attn_ext API
* ggml : fix GQA support in ggml_flash_attn_ext
* ggml : online attention (CPU)
* metal : initial implementation
* metal : f16 precision
* metal : reduce branches
* metal : specialize for head size
* wip : 8 rows per simd group
* wip : 4 rows per simd group
* wip : template for rows per warp
* metal : parallelize across KV size
* metal : parallel reduce across heads
* metal : efficient flash_attn_f16 implementation
* metal : avoid redundant loads of the attention
* metal : scale and mask in matrix form
* metal : fix comment
* llama : avoid ggml_cast, use F32 query
* metal : add parallel reduce version (disabled)
* metal : move output into local memory + optimize
- the result from each simdgroup now stays in the registers
- significantly reduced SRAM usage
- more efficient skipping of -INF blocks
- avoid simdgroup barrier in hot loop
- add comments
* metal : add tests, fix scaling, support C > 32
* metal : improve precision
* ggml : fix f16 mad
* metal : minor
* metal : support Q > 8
* tests : add ATTN tests
* metal : disable buffer allocation logs
* tests : more
* metal : faster inner loop for C == 32
* metal : fix array initialization
* tests : ifdef
* ggml : switch to padded F16 mask for ggml_soft_max, ggml_flash_attn_ext
* ggml : fix ggml_soft_max mask requirement
* cuda : fix soft_max to use correct mask size
* cuda : add flash_attn kernel (wip)
* metal : optimize softmax for C > 32
* metal : optimize softmax
* tests : minor fix
* cuda : avoid zeroing fragments
* tests : update dims
* cuda : fix __hisinf() result check
* cuda : avoid warp_reduce for smax
* cuda : use int instead of int64_t
Noticeably improves performance (thanks to Johannes)
* cuda : make loops use the same loop values
Thanks Johannes again for the tip
* cuda : unroll some of the loops
* cuda : avoid __hisinf branches
* cuda : use half2 in softmax
* cuda : switch to 1 warp for bs > 16
* cuda : speed-up reduce part of the kernel
* cuda : unroll Q*K^T loop
* cuda : fix -INF block check
* cuda : simplify softmax
* cuda : fix matrix names
* cuda : minor
* llama : adapt to F16 KQ_pos
* llama : adapt new models to F16 KQ_mask
* ggml : fix F16 store (ARM NEON)
* llama : fix type of KQ_mask and KQ_pos
* ggml : fix CPU soft_max
* tests : add hs=256
* cuda : fix build
* metal : improve perf via smaller int registers
* cuda : adapt soft_max to F16 mask and pos
* CUDA: faster FlashAttention, kernel for bs == 1
* 16 cols for Phi-2
* no vec for hs, no hs==256 ncols==32 for Volta
* adjust kernel selection logic
* 4 warps, 256 stride for all D
* no ncols == 64
* Multiple parallel blocks for batch size 1
* fix compile warnings
* fix excessive KQ_b loads
* fix cmake build
* fix KV cache padding, NaN from INFINITY (#6438)
* llama : flash_attn cparam + fix defrag
* server: support flash_attn param
* server: bench: enable flash_attn param
* CUDA: refactor host code, dyn. par. blocks
* fix flash_attn_vec_f16 race condition
* flush softmax exp below threshold to 0
* store temp KQ in registers
* Calculate KQ as FP32 if KQV has GGML_PREC_F32
* Add __hgt2_mask implementation for CUDA 11
* fix KQ FP32 precision fpr parallel_blocks > 1
* llama-bench : add -fa,--flash-attn arg
* metal : add BS=1 kernel for flash attention (#6508)
* metal : add BS=1 kernel for flash attention (wip)
* metal : support more than 1 warps
* metal : opts
* metal : opt
* metal : switch to parallel reduce
* metal : reduce registers
* metal : simplify
* metal : initial FA vec kernel
* metal : use F32 attention accumulators
* batched-bench : add fattn arg
* llama : simplify llama_build_kv_store
ggml-ci
* llama : adapt build_olmo to changes
* ggml : fix arm fp16 store on windows
* metal : clean-up
* metal : clean-up kernel code
* metal : minor
* tests : remove benchmarks
ggml-ci
* ggml : fix avx512 const correctness
ggml-ci
* ggml : fix soft_max with bias on CPU
ggml-ci
* common : print --flash-attn in help
* ggml : fix num dimensions in ggml_flash_attn_ext
* llama : force disable flash attention for incompatible models
* ggml : ggml_soft_max support F16/F32 mask/pos
ggml-ci
* cuda : uint -> uint32_t
* cuda : "constexpr dim3" -> "const dim3"
ggml-ci
* cuda : try to fix __hgt2_mask
ggml-ci
* ggml : add TODO's for F16/F32 mask/pos support in other backends
* llama : replace bool need_kq_pos with use_alibi
* llama : prep ALiBi support for BERT models
ggml-ci
* llama : fix n_batch requirements
ggml-ci
* cont
* server : add help for --flash-attn arg
* llama : disable FA for AMD
* tests : remove TMP_ATTN_BENCH
ggml-ci
* llama : support save/load state with FA enabled
ggml-ci
* ci : add CUDA save-load-state tests
ggml-ci
* llama : llama_kv_cache_clear zeroes data + fix save-load seq
ggml-ci
* llama : fix copy-paste errors, add TODO
* llama : disallow incompatible states
* llama : update llama_state_get_size after v_trans field
* metal : remove tmp log
* llama : add static reminder for llama_state_get_size
* metal : fix max nsg
ggml-ci
* ci : fix arg order
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Co-authored-by: Pierrick HYMBERT <pierrick.hymbert@gmail.com>
* imatrix: save the dataset file used in the output file
* llama: support kv overrides type string string
* common: factorize KV Overrides parsing between common and server
* quantize: add imatrix n entries and dataset KV metadata
quantize: factorize KV Overrides parsing between common
#6656
* llama: remove kv override str_value initialization as it does not compile on some toolchain
* quantize: add imatrix m_last_call as `quantize.imatrix.chunks_count`
* quantize: add imatrix filename in KV
* llama: add llama_model_kv_override_free
* common: add llama_model_kv_override_free
common: free kv override if used after model loading
* llama: finally move the string KV override value to the stack
* llama : minor
* no need to add a NUL to the std::vector, std::string can be initialized from a pair of iterators.
Co-authored-by: slaren <slarengh@gmail.com>
* kv override: ensure string termination
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
* server: cap n_predict if not set to n_ctx_train
* server: fix infinite loop
* server: infinite loop, move in process_token
server: infinite loop: set stop limit to true
* minor: spaces
* minor: spaces
* server: include prompt tokens in the EOS limit
* add support for moondream vision language model
This required making the following changes to the CLIP model:
1. Support for patch embedding bias.
2. Make class embedding and pre-layernorm optional.
3. Add support for post-layernorm.
* Update examples/llava/clip.cpp
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit renamesthe lerp (linear interpolation) function in clip.cpp
to avoid a conflict with the lerp function in the <cmath> standard C++
library when using c++20.
The motivation for this change is to enable projects that use c++20 to
be able to compile clip.cpp without having to resort to patching it. The
lerp function was added to cmath in version C++20 (202002L) and is why
this is not causing any issue at the moment as C++11/C++17 is currently
used by llama.cpp.
I realize that llama.cpp uses either C++11 (or C++17 in the case for
SYCL) but wanted to ask if this would be an acceptable change just the
same.
Refs: https://en.cppreference.com/w/cpp/numeric/lerp
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Implement '--keep-split' to quantize model into several shards
* Add test script
* Update examples/quantize/quantize.cpp
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Split model correctly even if tensor id is out-of-order
* Update llama_model_quantize_params
* Fix preci failures
---------
Co-authored-by: z5269887 <z5269887@unsw.edu.au>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* fix: revert showing control tokens by default
* feat: revert changes to default behavior of llama_token_to_piece; provide overridden declaration to receive "bool special" param to toggle showing control tokens
* feat: use the overridden declaration of llama_token_to_piece from common/common.cpp to specify "false" so that control tokens are not shown in chat completion responses"
* common : simplify
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* `build`: generate hex dumps of server assets on the fly
* build: workaround lack of -n on gnu xxd
* build: don't use xxd in cmake
* build: don't call xxd from build.zig
* build: more idiomatic hexing
* build: don't use xxd in Makefile (od hackery instead)
* build: avoid exceeding max cmd line limit in makefile hex dump
* build: hex dump assets at cmake build time (not config time)
* Support Llama 3 conversion
The tokenizer is BPE.
* style
* Accept suggestion
Co-authored-by: Sourab Mangrulkar <13534540+pacman100@users.noreply.github.com>
* llama : add llama_token_is_eog()
ggml-ci
* llama : auto-detect more EOT tokens when missing in KV data
* convert : replacing EOS token is a hack
* llama : fix codegemma EOT token + add TODOs
* llama : fix model type string for 8B model
---------
Co-authored-by: Sourab Mangrulkar <13534540+pacman100@users.noreply.github.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ggml : group all experts in a single ggml_mul_mat_id
cuda : improve mmid row copy
* cuda : fix bin bcast with non-cont src0
* test-backend-ops : only run all mul mat tests for base types
* llama : disable moe offloading with SYCL
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This change upstreams llamafile's cpu matrix multiplication kernels
which improve image and prompt evaluation speed. For starters, Q4_0
and Q8_0 weights should go ~40% faster on CPU. The biggest benefits
are with data types like f16 / f32, which process prompts 2x faster
thus making them faster than quantized data types for prompt evals.
This change also introduces bona fide AVX512 support since tinyBLAS
is able to exploit the larger register file. For example, on my CPU
llama.cpp llava-cli processes an image prompt at 305 tokens/second,
using the Q4_K and Q4_0 types, which has always been faster than if
we used f16 LLaVA weights, which at HEAD go 188 tokens/second. With
this change, f16 LLaVA performance leap frogs to 464 tokens/second.
On Intel Core i9-14900K this change improves F16 prompt perf by 5x.
For example, using llama.cpp at HEAD with Mistral 7b f16 to process
a 215 token prompt will go 13 tok/sec. This change has fixes making
it go 52 tok/sec. It's mostly thanks to my vectorized outer product
kernels but also because I added support for correctly counting the
number of cores on Alderlake, so the default thread count discounts
Intel's new efficiency cores. Only Linux right now can count cores.
This work was sponsored by Mozilla who's given permission to change
the license of this code from Apache 2.0 to MIT. To read more about
what's improved, and how it works, see: https://justine.lol/matmul/
This commit updates the hf.sh script usage to include the --outdir option
and specifies the models directory as the output directory.
The motivation for this is to avoid cluttering the root directory with
model files.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* Fix --split-max-size
Byte size calculation was done on int and overflowed.
* add tests.sh
* add examples test scripts to ci run
Will autodiscover examples/*/tests.sh scripts and run them.
* move WORK_PATH to a subdirectory
* clean up before and after test
* explicitly define which scripts to run
* add --split-max-size to readme
* disable mmap to fix memcpy crash, add missed cmd in guide, fix softmax
* refactor to disable mmap for SYCL backend
* fix compile error in other os
* refactor the solution, use host buf to fix it, instead of disable mmap
* keep to support mmap()
* use host buff to reduce malloc times
* revert to malloc/free solution, for threaad safe
* infill : add download instructions for model
This commit adds instructions on how to download a CodeLlama model
using the `hf.sh` script. This will download the model and place it
in the `models` directory which is the same model use later by the
infill example.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* squash! infill : add download instructions for model
Clarify the reason for using CodeLlama.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* gguf-debug: Example how to use ggml callback for debugging
* gguf-debug: no mutex, verify type, fix stride.
* llama: cv eval: move cb eval field in common gpt_params
* ggml_debug: use common gpt_params to pass cb eval.
Fix get tensor SIGV random.
* ggml_debug: ci: add tests
* ggml_debug: EOL in CMakeLists.txt
* ggml_debug: Remove unused param n_batch, no batching here
* ggml_debug: fix trailing spaces
* ggml_debug: fix trailing spaces
* common: fix cb_eval and user data not initialized
* ci: build revert label
* ggml_debug: add main test label
* doc: add a model: add a link to ggml-debug
* ggml-debug: add to make toolchain
* ggml-debug: tests add the main label
* ggml-debug: ci add test curl label
* common: allow the warmup to be disabled in llama_init_from_gpt_params
* ci: add curl test
* ggml-debug: better tensor type support
* gitignore : ggml-debug
* ggml-debug: printing also the sum of each tensor
* ggml-debug: remove block size
* eval-callback: renamed from ggml-debug
* eval-callback: fix make toolchain
---------
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit adds an option to the gguf example to not check the tensor
data.
The motivation for this is that it can be nice to use the gguf tool to
read other .gguf files that were not created by the gguf tool.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
Key changes:
* BERT conversion: fix abuse of LlamaHfVocab, do not set BOS or EOS
* Nomic Embed conversion: pad vocab instead of slicing embedding tensor
* llama_tokenize: handle added special tokens like HF does
* llama : save and restore kv cache for single seq id
* remove trailing whitespace
* respond error in case there's no space in the kv cache
* add kv seq save restore to test case
* add --slot-save-path arg to enable save restore and restrict save location
* Returning 0 for some cases, instead of asserting.
* cleanup error cases
* rename sequence state functions
* rename state get set functions
* add previous function names back in with DEPRECATED notice
* update doc
* adjust endpoints to preferred style
* fix restoring zero cell count
* handle seq rm return value
* unused param
* keep in the size check
* fix return types
* add server test case for slot save restore
* cleanup
* add cake
* cleanup style
* add special
* removing a whole sequence never fails
* move sequence state file functionality from server to llama to match session api and add version tags
* catch exceptions on save as well
* error log messages
* check types for stricter restore
* update server doc
* readme : update API changes date
* strict filename validation
* move include, reject bom as well
* also reject empty filename
* reject whitespace and trailing dot
---------
Co-authored-by: Martin Evans <martindevans@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ci: bench: support sse and fix prompt processing time
server: add tokens usage in stream mode
* ci: bench: README.md EOL
* ci: bench: remove total pp and tg as it is not accurate
* ci: bench: fix case when there is no token generated
* ci: bench: change to the 95 percentile for pp and tg as it is closer to what the server exports in metrics
* ci: bench: fix finish reason rate
* ci: bench: change trigger path to not spawn on each PR
* ci: bench: add more file type for phi-2: q8_0 and f16.
- do not show the comment by default
* ci: bench: add seed parameter in k6 script
* ci: bench: artefact name perf job
* Add iteration in the commit status, reduce again the autocomment
* ci: bench: add per slot metric in the commit status
* Fix trailing spaces
* Typo fix to server's README.md
Fix minor typo ("tonen") in server README.
* server readme grammar/style fixes.
Quickly went through this file to look for inconsistencies in
presentation of defaults, flag options, and looked for typos
and grammar issues.
Not perfect, but hopefully improved.
* Update README.md
Remove an extra space before newline.
* ggml : update mul_mat_id to use the same tensor for all the experts
* update cuda
* minor
* update metal
* update test-backend-ops
* fix cuda
* Update ggml-metal.m
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* update convert.py
* update convert-hf-to-gguf.py
* update convert.py for mixtral hf models
* Update convert-hf-to-gguf.py
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* cuda : support non-pow-2 number of experts
* allow quantize to work for split and merged experts models in the same way
* cleanup + disable mmap automatically with split tensors models
* update imatrix
* test-backend-ops : test qwen argsort
* update grok model loading
* llama : add merged experts tensors to the grok tensor map
* minor
* gguf : bump version
* fix quantizing of merged experts
* convert-hf-to-gguf.py : update grok (untested)
* make linter happy
* cuda/argsort : use shared memory instead of pool memory
* convert : fix grok tensor names
* metal : add support for non-pow-2 argsort
* llama : more loader cleanup, better error checking
* cuda : fix warning
* llama : still use mmap for loading old models, but copy the data to a host buffer
* add review note
* llama : remove ffn tensor counting + add sanity check
ggml-ci
* convert : fix handling of n_experts == None
ggml-ci
* imatrix : fix ncall counters
* llama : produce error if imatrix size does not match
* quantize : terminate on errors + trace logs
ggml-ci
* metal : pad shared memory to 16 bytes
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* embedding : show full embedding for single prompt
To support the use case of creating an embedding for a given prompt, the entire embedding and not just the first part needed to be printed.
Also, show cosine similarity matrix only if there is more than one prompt, as the cosine similarity matrix for a single prompt is always `1.00`.
* Update examples/embedding/embedding.cpp
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama : greatly reduce logits memory usage
* llama : more compact state saving and reloading
* llama : fix lctx.n_outputs not being set before building graph
* perplexity : adapt to the logits API changes
* perplexity : fix Winogrande, use correct logits for second choice start
The first logits used to evaluate the second choice were not from
the end of the common prefix; instead, they were the logits from the end
of the first choice. This has been corrected.
The previous implementation sometimes had outliers in the scores of
choices for some tasks, and the logic to skip choices words
in the log-likelihood evaluation probably was an attempt to reduce those,
but it was complex and didn't quite seem to be the right thing.
This is simpler now, and the outlier scores aren't there anymore.
* perplexity : normalize spaces and punctuation in Winogrande sentences
* llama : fix embedding conditions
* llama : fix llama_get_embeddings_ith when the resulting id is 0
* llama : fix wrong n_outputs in llama_set_inputs
A mismatch happened when using a smaller n_ubatch than n_batch and then using
llama_batch_get_one(). The decision of what n_outputs should be now almost
fully depends on how lctx.n_outputs is set in llama_decode_internal.
The conditions are simpler this way.
* llama : when saving the state, recalculate n_outputs
This ensures the correct number of outputs for the entire previous batch
is stored in the session file, even when n_ubatch is smaller than n_batch.
* llama : fix not-skipping outputs of non-causal models
* llama : fix running a batch with n_outputs == 0
It previously worked because lctx.inp_out_ids was not initialized,
so it pointed to some garbage address which was somehow still valid when I
ran my tests.
* llama : keep same graph topology even when n_outputs == 0
* ggml : saner ggml_can_repeat with empty tensors
* ggml : future-proof ggml_is_empty by using GGML_MAX_DIMS - 1
* ggml : do not multi-thread ops returning empty tensors
* ggml : make ggml_is_empty public and work with views
* llama : use a vector for ctx->output_ids
* llama : rework reallocation logic for llama_output_reserve
Now comparing the actual size with the new total size of the output buffer
to allow more efficient enabling and disabling of the embeddings
and/or logits output in the future.
* ggml : skip empty tensors in all backends
* llama : fix llama_output_reserve nullptr deref when new_size is 0
* perplexity : make Winogrande work as it does on master
The problems with the Winogrande implementation will
need to be fixed in a separate PR to ease review.
* llama : clearer error messages for invalid logits or embeddings ids
* llama : assert all models that can have inp_out_ids
Since the graph topology is now constant, this presence check
can be done even when there are no outputs.
* llama : assert logits and embd buffers exist before writing to them
* llama : handle errors from llama_output_reserve at call sites
* perplexity : make hellaswag and multiple-choice outputs identical to master
Due to how the KV cache is updated, the logprobs for tokens in a batch
are very slightly affected by the other tokens present in the batch,
so to make hellaswag and multiple-choice return exactly the same results
as on master, the last token of each sequence needs to be evaluated
even though its output is not used at all.
This will probably be changed back in the future to make these benchmarks
a tiny bit faster.
* perplexity : fix division by zero when using less than 100 multiple-choice tasks
* llama : allow loading state saved with a different ctx size
When loading a session file, the context size is now only required to be
at least enough to load the KV cells contained in that session file,
instead of requiring to use exactly the same context size as when saving.
Doing this enables the use-case of extending or shrinking the context size
of a saved session.
This breaks existing session files because the meaning of kv_buf_size
is slightly changed (previously it was the size of the whole KV cache,
now it's only the size of the saved part of it). This allows for
finer-grained sanity checks when loading in an effort to keep kv_buf_size
useful even when the kv_size is changed.
* llama : minor
ggml-ci
* readme : update recent API changes, and warn about Vulkan
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* iq1_m: basics
* iq1_m: basics-2
* iq1_m: CUDA dequantize works
Very 1st shot I get PPL = 9.76 for LLaMA-v2-7B.
* iq1_m: separate shifts for each group of 8 in a block
We get
PPL(LLaMA-v2-7B ) = 9.2810
PPL(LLaMA-v2-13B) = 6.8105
Not bad, but slightly higher than
sqrt(PPL(IQ1_S) * PPL(IQ2_XXS))
which is the expected outcome given that IQ1_M is
halfway between IQ1_S and IQ2_XXS in terms of bpw.
From this, we would expect
PPL = 9.14 for LLaMA-v2-7B
PPL = 6.63 for LLaMA-v2-13B
* iq1_m: go to 3-bit scales
There is slight increase in PPL, but the 0.0625 bpw reduction
in size is totally worth it.
We now have
PPL(LLaMA-v2-7B ) = 9.4469 at 1.96 bpw
PPL(LLaMA-v2-13B) = 6.8717 at 1.93 bpw
PPL(LLaMA-v2-70B) = 4.8568 at 1.85 bpw
* iq1_m: scalar dot product
* iq1_m: AVX2 dot product
* iq1_m: very slightly faster AVX2 dot product
* iq1_m: ARM_NEON dot product
Works, but very slow (10.5 t/s)
* iq1_m: Metal - dequantize works, dot product does not
* iq1_m: Metal now works
About the same performance as iq1_s.
* iq1_m: minor
* iq1_m: checking pure iq1_m quantization
It is pretty bad: PPL(LLaMA-v2-7B) = 34 if we quantize output.weight
with Q4_K.
* iiq1_m: slightly faster ARM_NEON dot product
10.5 t/s -> 11.65 t/s
* iq1_m: faster ARM_NEON dot product
11.65 t/s -> 14.9 t/s
* iq1_m: another minor ARM_NEON dot product improvement
14.9 -> 15.0 t/s
* iq1_m: small PPL improvement via super-block scale adjustment
After quantizing block scales redo the super-block scale fit.
PPL(LLaMA-v2-7B ) = 9.3346
PPL(LLaMA-v2-13B) = 6.8419
PPL(LLaMA-v2-70B) = 4.8294
PPL(Mistral-7B ) = 8.1624
* iq1_m: adapt to CUDA refactoring
* iq1_m: remove unused variable
We have progressed to warnings being errors.
* iq1_m: add to backend-ops tests
* iq1_m: fix Windows ARM
* iq1_m: use common definition of iq1m_scale_t
* cuda: assert -> NO_DEVICE_CODE
* iq1_M: PR comments
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* quantize: be able to override metadata by key
* minor : spacing
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* embedding: assign `n_ubatch` value, print error on `n_batch` overflow
* Update examples/embedding/embedding.cpp
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
* use %ld instead of %lld
* Revert "use %ld instead of %lld"
This reverts commit ea753ede90a86a0699f65878cc8e2020ff5eabb8.
---------
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
* sampling: remove duplicated code for probability distribution access
* free original_logits
* fix original_logits allocation
* fixes based on review @cebtenzzre
* change function name to `llama_sampling_prepare`
* llama: llama_split_prefix fix strncpy does not include string termination
common: llama_load_model_from_url:
- fix header name case sensitive
- support downloading additional split in parallel
- hide password in url
* common: EOL EOF
* common: remove redundant LLAMA_CURL_MAX_PATH_LENGTH definition
* common: change max url max length
* common: minor comment
* server: support HF URL options
* llama: llama_model_loader fix log
* common: use a constant for max url length
* common: clean up curl if file cannot be loaded in gguf
* server: tests: add split tests, and HF options params
* common: move llama_download_hide_password_in_url inside llama_download_file as a lambda
* server: tests: enable back Release test on PR
* spacing
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* spacing
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* spacing
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* quantize: be able to specify the output tensor type
* quantize: be able to specify the token embedding tensor type
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* split: support in llama_model_loader
* avoid copying the entire vector
Co-authored-by: slaren <slarengh@gmail.com>
* split: move llama_tensor_offset to llama_model_loader
* llama_model_loader: PR feedbacks:
- use only one gguf_context for metadata only
- store all ggml_context in a vector as the files and mappings
- store all weights in a vector along with the source tensor
- rename ctx_gguf to meta
- rename ctx_meta to contexts
* avoid copying the entire vector
* Simplify this by making these optional, switch some layer creation tensor optional
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Handle optional tensors
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama_model_loader: fail if backend cannot allocate buffer
* fix mmap buffer management
* llama_model_loader: map file to backend buffer if the allocation succeeds only
* llama_model_loader: only map tensors included in the context
* llama_model_loader: minor, use same variable name for consistency, fix spacing in types cast
* llama_model_loader: fail if any of backend buffer cannot be allocated
* spacing
Co-authored-by: slaren <slarengh@gmail.com>
* fix loop over pointer
Co-authored-by: slaren <slarengh@gmail.com>
* llama_model_loader: if n_tensors declared not equals to loaded tensors in split, throw an exception instead of asserting
* llama_model_loader: ensure mappings vector has the expected size
* llama_model_loader: use at instead of operator[] if this should never add to the map.
* llama_model_loader: immediately add the backend buffer to the model buffers in order to free them if an error occurs in the next allocation. Reserve the expected size.
* llama_model_loader: be sure the model mappings has enough capacity before allocating backend buffer
* llama_model_loader: fix map -> unordered map
* llama_split_prefix: use a clearer version, not pass split path len but dest max len.
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
* llama : minor
ggml-ci
* llama : introduce some typedef helpers
* docs: add model shard in hot topic
* llama_model_loader: put mapping in a unique_ptr from the moment it is allocated
Co-authored-by: slaren <slarengh@gmail.com>
* fix llama_split_prefix
---------
Co-authored-by: slaren <slarengh@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Xuan Son Nguyen <thichthat@gmail.com>
* k_cache: be able to use Q5_0
* k_cache: be able to use Q5_1 on CODA
* k_cache: be able to use Q5_0 on Metal
* k_cache: be able to use Q5_1 on Metal
* k_cache: be able to use IQ4_NL - just CUDA for now
* k_cache: be able to use IQ4_NL on Metal
* k_cache: add newly added supported types to llama-bench and CUDA supports_op
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Kawrakow
Andrew Chan
Nexes the Elder
saood06
Kawrakow
Douglas Hanley
Shuichi Tsutsumi
luoyu-intel
sasha0552
Sigbjørn Skjæret
Calvin Laurenson
Xuan Son Nguyen
Radoslav Gerganov
slaren
Olivier Chafik
Georgi Gerganov
Johannes Gäßler
mgroeber9110
Christian Zhou-Zheng
woodx
arch-btw
nickp27
Yazan Agha-Schrader
HanishKVC
Galunid
Nathan Epstein
Ikko Eltociear Ashimine
Brian
Justine Tunney
Elton Kola
Mikko Juola
Neo Zhang
liuwei-git
Amir
jaime-m-p
Fred Douglas
strawberrymelonpanda
Leon Knauer
Pierrick Hymbert
Vaibhav Srivastav
dm4
Steve Grubb
Ryuei
k.h.lai
Benjamin Findley
Joan Fontanals
Andrei
Ahmet Zeer
compilade
JohnnyB
Dawid Potocki
Johan
jukofyork
Kyle Mistele
RhinoDevel
omahs
kunnis
maor-ps
l3utterfly
Daniel Bevenius
cpumaxx
vik
jiez
Pedro Cuenca
Jan Boon
Austin
Neo Zhang Jianyu
Rene Leonhardt
Clint Herron
Ralph Soika
Jared Van Bortel
Ed Lee
Ting Sun
Shakhar Dasgupta
Xiao-Yong Jin
Fattire
JH23X
Ziang Wu
Eric Zhang
howlger
compilade
Minsoo Cheong
fraxy-v
Jan Boon