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>.
* 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>
* add basic tensor data validation function
* add --check-tensors command line argument
tensor validation is disabled by default and can be enabled by adding
`--check-tensors` to the command line arguments.
quantize always validates tensors.
* 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>
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/
* 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>
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_sampling_sample with default args is more naively usable
* Batches populated by either llama_batch_get_one or llama_batch_add work with default args
* Previously get_one could use the default argument
* Previously add should usually have used the last index where logits[idx] == true
* This hopefully encourages the use of llama_batch_add
* By giving expected results when using default arguments.
* Adds "negative indexing" feature to llama_get_logits_ith and llama_get_embeddings_ith
* Believed to work with any currently well behaved program
* Default arg now works for both cases (previously would give strange results for add case)
* Any non-negative number is unaffected and behaves as previously
* Negative arguments were previously invalid.
* Implemented as a special case of indexing as suggested by @compilade in https://github.com/ggerganov/llama.cpp/pull/6519
* Fixed mismatch type errors
* cited in macOS CI tests
* Missed in original updates based on PR feedback in https://github.com/ggerganov/llama.cpp/pull/6519
* 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>
This commit removes one of the two identical checks for curl being NULL
in llama_load_model_from_url.
Signed-off-by: Daniel Bevenius <daniel.bevenius@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>
* 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>
* control vector api and implementation
* control-vectors : minor code style updates
* disable control vector when data == nullptr
use -1 for disabled range (also on init) in case we ever support controlling layer 0 (embeddings)
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama : add pipeline parallelism support for batch processing with multiple CUDA GPUs
ggml-ci
* server : add -ub, --ubatch-size parameter
* fix server embedding test
* llama : fix Mamba inference for pipeline parallelism
Tested to work correctly with both `main` and `parallel` examples.
* llama : limit max batch size to n_batch
* add LLAMA_SCHED_MAX_COPIES to configure the number of input copies for pipeline parallelism
default increase to 4 (from 2)
changing this value may improve performance for some systems, but increases memory usage
* fix hip build
* fix sycl build (disable cpy_tensor_async)
* fix hip build
* llama : limit n_batch and n_ubatch to n_ctx during context creation
* llama : fix norm backend
* batched-bench : sync after decode
* swiftui : sync after decode
* ggml : allow ggml_get_rows to use multiple threads if they are available
* check n_ubatch >= n_tokens with non-casual attention
* llama : do not limit n_batch to n_ctx with non-casual attn
* server : construct batch with size of llama_n_batch
* ggml_backend_cpu_graph_compute : fix return value when alloc fails
* llama : better n_batch and n_ubatch comment
* fix merge
* small fix
* reduce default n_batch to 2048
---------
Co-authored-by: Francis Couture-Harpin <git@compilade.net>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* mamba : begin working on support for Mamba SSM
* mamba : begin figuring out how to (ab)use the kv cache for Mamba
* mamba : recurrent inference almost works, but incoherent
* mamba : recurrent inference WORKS!!!
* convert : optionally use d_conv and d_state from config.json for Mamba
* mamba : refactor recurrent conv, resulting in 20% perf increase
It's still slower than I'd like, but I did not really optimize `ggml_exp` yet.
I also refactored `ggml_exp` to work with tensors with more than 2 dimensions.
* ggml : parallelize ggml_exp
This results in 8% faster token generation for Mamba-130M.
* mamba : simplify the conv step with a self-overlapping view
Turns out the conv_state can be made smaller by one column.
Note that this breaks existing GGUFs of Mamba,
because the key_value_length field is tied to the conv_state size.
Convolution with a self-overlapping view is cool!
And it's much simpler than what I initially thought would be necessary
to make the convolution step work with more than 1 token at a time.
Next step is to make the SSM step work on batches of tokens too,
and thus I need to figure out a way to make a parallel selective scan
which will keep the ssm_state small and won't make it bigger
by a factor of (n_layer * batch_size).
* llama : fix Mamba KV self size wrongly displaying as f16 instead of f32
Relatedly, I also tried to see if other types than f32 worked for the states,
but they don't, because of the operators used.
It's probably better anyway to keep lots of precision there,
since the states are small anyway.
* mamba : fix self-overlapping view depth stride
* mamba : handle batches of more than 1 token
This means running Mamba no longer crashes when using the default settings!
And probably also slightly faster prompt processing.
Both batched and non-batched processing yield the same output.
Previously, the state was not cleared when starting a sequence.
Next step is to make the KV cache API work as expected for Mamba models.
* ggml: add ggml_ssm_scan to help with parallel selective scan
If the selective scan was implemented without a custom operator,
there would be waaay too many nodes in the graph. For example,
for Mamba-130M, with a batch size of 512 (the default),
a naive selective scan could add at least 24*512=12288 nodes,
which is more than LLAMA_MAX_NODES (8192),
and that's only for the smallest Mamba model.
So it's much cleaner with a custom operator.
Not sure about the name, though.
* ggml : in ggml_ssm_scan, merge multiple rows in the same vec operation
This will help with performance on CPU if ggml_vec_mul_f32
and ggml_vec_add_f32 are ever optimized with SIMD.
* mamba : very basic quantization support
Mostly works, but there is currently no difference
between the variants of a k-quant (e.g. Q4_K_S and Q4_K_M are the same).
Most of the SSM-specific weights can be kept in f32 without affecting
the size that much, since they are relatively small.
(the linear projection weights are responsible for most of Mamba's size)
Too much quantization seems to make the state degrade quite fast, and
the model begins to output gibberish.
It seems to affect bigger models to a lesser extent than small models,
but I'm not sure by how much.
Experimentation will be needed to figure out which weights are more important
for the _M (and _L?) variants of k-quants for Mamba.
* convert : fix wrong name for layer norm weight of offical Mamba models
I was using Q-bert/Mamba-* models before, which have a slighlty different
naming scheme for the weights.
(they start with "model.layers" instead of "backbone.layers")
* mamba : fuse more steps of the SSM scan in the ggml_ssm_scan operator
This increases performance on CPU by around 30% for prompt processing,
and by around 20% for text generation.
However, it also makes the ggml_exp and ggml_soft_plus operators unused.
Whether or not they should be kept will be decided later.
* convert : for Mamba, also consider the "MambaLMHeadModel" arch name
It's the name of the class of the official implementation,
though they don't use it (yet) in the "architectures" field of config.json
* mamba : fix vocab size problems with official models
The perplexity was waaaay to high for models with a non-round vocab size.
Not sure why, but it needed to be fixed in the metadata.
Note that this breaks existing GGUF-converted Mamba models,
but **only if** the vocab size was not already rounded.
* ggml : remove ggml_exp and ggml_soft_plus
They did not exist anyway outside of this branch,
and since ggml_ssm_scan fused operations together, they are unused.
It's always possible to bring them back if needed.
* mamba : remove some useless comments
No code change.
* convert : fix flake8 linter errors
* mamba : apply suggestions from code review
* mamba : remove unecessary branch for row-wise ssm_state and C multiplication
It was previously done to avoid permuting when only one token is processed
at a time (like when generating text), but permuting is cheap,
and dynamically changing the compute graph is not future-proof.
* ggml : in ggml_ssm_scan, use more appropriate asserts
* ggml : rename the destination pointer in ggml_compute_forward_ssm_scan_f32
* mamba : multiple sequences, but one at a time
This is a step towards making this Mamba implementation usable
with the server example (the way the system prompt is kept when clearing
the client slots will need to be changed before this can work, though).
The KV cache size for this kind of model is tied to the maximum number
of sequences kept at any single time.
For now, this number is obtained from n_parallel (plus one,
to have an extra sequence to dedicate to the system prompt),
but there might be a better way to do this which won't also
make the main example use 2 cells even if only 1 is really used.
(for this specific case, --parallel 0 helps)
Simultaneous sequence processing will probably require changes to
ggml_ssm_scan, and possibly a new operator for the conv step.
* mamba : support llama_kv_cache_seq_cp
This (mis)uses the logic around K shifts, because tokens in a state
can't be shifted anyway, and because inp_K_shift has the right shape and type.
Using ggml_get_rows is a nice way to do copies, but copy chains can't work.
Fortunately, copy chains don't really seem to be used in the examples.
Each KV cell is dedicated to the sequence ID corresponding to its own index.
* mamba : use a state mask
It's cleaner than the previous heuristic of
checking for the pos of the first token in the batch.
inp_KQ_mask could not be re-used for this, because it has the wrong shape
and because it seems more suited to the next step of
simultaneous sequence processing (helping with the problem of
remembering which token belongs to which sequence(s)/state(s)).
* llama : replace the usage of n_ctx with kv_self.size in many places
* mamba : use n_tokens directly instead of n_tok
* mamba : in comments, properly refer to KV cells instead of slots
* mamba : reduce memory usage of ggml_ssm_scan
From 290.37 MiB to 140.68 MiB of CPU compute buffer size
with Mamba 3B with a batch size of 512.
The result tensor of ggml_ssm_scan was previously a big part
of the CPU compute buffer size. To make it smaller,
it does not contain the intermediate ssm states anymore.
Both y and the last ssm state are combined in the result tensor,
because it seems only a single tensor can be returned by an operator
with the way the graph is built.
* mamba : simultaneous sequence processing
A batch can now contain tokens from multiple sequences.
This is necessary for at least the parallel example, the server example,
and the HellaSwag test in the perplexity example.
However, for this to be useful, uses of llama_kv_cache_seq_rm/cp
will need to be changed to work on whole sequences.
* ggml : add ggml_ssm_conv as a new operator for the conv step of Mamba
This operator makes it possible to use and update the correct states
for each token of the batch in the same way as ggml_ssm_scan.
Other solutions which use existing operators would need loops which would
add too many nodes to the graph (at least the ones I thought of).
Using this operator further reduces the size of the CPU compute buffer
from 140.68 MiB to 103.20 MiB with Mamba 3B with a batch size of 512.
And (at least on CPU), it's a bit faster than before.
Note that "ggml_ssm_conv" is probably not the most appropriate name,
and it could be changed if a better one is found.
* llama : add inp_s_seq as a new input tensor
The most convenient implementation to select the correct state (for Mamba)
for each token is to directly get the correct index from a tensor.
This is why inp_s_seq is storing int32_t and not floats.
The other, less convenient way to select the correct state would be
to have inp_KQ_mask contain 1.0f for each state used by a token
and 0.0f otherwise. This complicates quickly fetching the first used
state of a token, and is also less efficient because a whole row
of the mask would always need to be read for each token.
Using indexes makes it easy to stop searching when there are
no more sequences for a token, and the first sequence assigned
is always very quickly available (it's the first element of each row).
* mamba : support llama_kv_cache_seq_cp copy chains
* mamba : support shifting and dividing the kv cache pos
* mamba : make the server and parallel examples work with whole sequences
A seq_id is dedicated to the system prompt in both cases.
* llama : make llama_kv_cache_seq_rm return whether it succeeded or not
* mamba : dedicate an input tensor for state copy indices
This is cleaner and makes it easier to adapt when/if token positions
(and by extension, inp_K_shift) are no longer integers.
* mamba : adapt perplexity, batched, and batched-bench examples
* perplexity : limit the max number of sequences
This adapts to what the loaded model can provide.
* llama : add llama_n_max_seq to get the upper limit for seq_ids
Used by the perplexity example.
* batched : pass n_parallel to the model's context params
This should have been there already, but it wasn't.
* batched-bench : reserve sequences to support Mamba
* batched-bench : fix tokens being put in wrong sequences
Generation quality isn't what's measured in there anyway,
but at least using the correct sequences avoids using non-consecutive
token positions.
* mamba : stop abusing attention metadata
This breaks existing converted-to-GGUF Mamba models,
but will allow supporting mixed architectures like MambaFormer
without needing to break Mamba models.
This will also allow changing the size of Mamba's states
without having to reconvert models in the future.
(e.g. using something else than d_conv - 1 columns for the conv_states
will not require breaking existing converted Mamba models again)
* gguf-py : add new KV metadata key-value pairs for Mamba
* llama : add new metadata key-value pairs for Mamba
* llama : guard against divisions by zero when n_head is 0
* mamba : rename "unlimited" KV cache property to "recurrent"
* mamba : more correctly update the "used" field of the KV cache
* ggml : in ggml_ssm_scan, use a threshold for soft_plus
This is how the official Mamba implementation does it,
and it's also what torch.nn.Softplus does.
* convert : for Mamba, fallback to internal NeoX tokenizer
The resulting models are exactly the same
as if the tokenizer.json and tokenizer_config.json of GPT-NeoX were there.
* mamba : support state saving and restoring
* ggml : implicitly pass src tensors through dst for Mamba-related ops
* mamba : clarify some comments
* server : fix cache_tokens not getting correctly resized
Otherwise, when the "we have to evaluate at least 1 token" special case
was triggered, an extra token was kept in cache_tokens even if it was
removed from the KV cache.
For Mamba, this caused useless prompt reprocessing when the previous
request triggered the above case.
* convert-hf : support new metadata keys for Mamba
For the models available at
https://huggingface.co/collections/state-spaces/transformers-compatible-mamba-65e7b40ab87e5297e45ae406
* mamba : rename metadata to be more similar to transformers library
This breaks existing converted-to-GGUF models,
but the metadata names are more "standard".
* mamba : support mamba-*-hf models
These models share their token_embd.weight with their output.weight
* mamba : add missing spaces
This is purely a formatting change.
* convert-hf : omit output.weight when identical with token_embd.weight
Only for Mamba for now, but it might be relevant for other models eventually.
Most Mamba models actually share these two tensors, albeit implicitly.
* readme : add Mamba to supported models, and add recent API changes
* mamba : move state_seq and state_mask views outside layer loop
A few tensors were also missing `struct` in front of `ggml_tensor`.
MSVC gives the following error with the existing macros:
`Error C2059 : syntax error: ','`
This patch adds `##` as a prefix to `__VA_ARGS__` to address this error.
* (WIP) Implement stochastic speculative decoding
* sample from residual distribution on draft accept failure
* fix#5657: force greedy sampling with probs when temp is 0
* remove p_accept parameter
* fix style
* remove unused variables
* add srand() in speculative.cpp
* replace use of rand() with mt19937 sampling
* fixes based on review (@JohannesGaessler)
* fix r random generation
* randomly select next sequence to verify + fix bug in memory freeing
* fix bug in active_seqs sync
* fix uniform int distribution initialization
* remove warnings from comparison between int and size_t
* check grammar in `llama_sample_probability_distribution_impl`
* remove malloc code by utilizing vectors
* add PR link to README
* suport multiple cards: split-mode - layer|row
* rm warning
* rebase with master, support tow new OPs, close feature for -sm=row, fix for unit test
* update news
* fix merge error
* update according to review comments
* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h
* Reverted Makefile
* Fixed include
* Removed sched.h from ggml.h, moved ggml_get_numa_affinity into ggml.c, removed trailing whitespace and fixed up a few inconsistent variables
* removed trailing whitespace
* Added numa options to allow finer grained control as well as plumbing for a new mirror mode that will require numa.h
* Reverting Makefile
* Fixed a number of issues with the move from BOOL to ggml_numa_strategies. Added a note about mirror mode note being implemented yet
* Removing MIRROR_MODE code for this PR
* Removing last bit of MIRROR_MODE code for this PR
* Removing unneeded branch in server.cpp example and moving get_numa_affinity and making it static
* Fixed lingering init_llama_backend() bool calls in tests and examples
* Remote enum llama_numa_strategies
* Revert bad merge with dynatemp flags
* add missing enum ggml_numa_strategies declaration and revert sync problem with master
* add missing enum ggml_numa_strategies declaration
* fixed ggml_init_numa variable
* Update ggml.h
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* Update READMEs with info about numa flags, change INTERLEAVE strategy name to DISTRIBUTE everywhere, implement the improved distribution strategy from @rankaiyx, fix a spelling mistake and un-merge some bad merges
* split numa init out from llama_backend_init and created llama_numa_init. Updated all code paths and samples
* Fix up some boolean vs enum comparisons
* Added #ifdefs for non-Linux OS that don't have cpu_set_t datatype
* Update ggml.h
Align enum values
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml.c
Remove whitespace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update ggml.c
align paremeters
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/server/server.cpp
remove whitespace and align brace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update common/common.cpp
Remove whitespace and align brace
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* unified ggml_numa_strategy enum and fixed text alignment in server.cpp example
* Update ggml.c
simplified return for platforms without NUMA support
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* removed redundant else from cli argument processing of --numa
* whitespace
---------
Co-authored-by: root <root@nenya.lothlorien.ca>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
* common: use enums for sampler types
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* minor : spaces
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* ggml: aarch64: implement smmla kernel for q8_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q8_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_1_q8_1 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_1_q8_1 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: update unit tests for the new vec_dot interface
* llama.cpp: add MATMUL_INT8 capability to system_info
* Initial Vulkan multi-gpu implementation
Move most global variables into backend context
* Add names to backend device functions
* Add further missing cleanup code
* Reduce code duplication in tensor split layer assignment
* generalize LLAMA_SPLIT_LAYER for all backends, do not expose device count and memory in llama.h
* Only do device info print in the beginning and initialize one backend for cpu assist
Add missing cleanup code
* Rework backend memory management to make sure devices and buffers get properly allocated and freed
* Rename cpu assist free function
---------
Co-authored-by: slaren <slarengh@gmail.com>
* Fix Vulkan F16 models
* Fix Vulkan context shift crash
* Add Vulkan to common.cpp dump_non_result_info_yaml function
* Fix bug in Vulkan CPY op
* Fix small matrix multiplication errors in AMD GPUs on Windows or with amdvlk
Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>
---------
Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>
* first update for migration
* update init_cublas
* add debug functio, commit all help code
* step 1
* step 2
* step3 add fp16, slower 31->28
* add GGML_LIST_DEVICE function
* step 5 format device and print
* step6, enhance error check, remove CUDA macro, enhance device id to fix none-zero id issue
* support main device is non-zero
* step7 add debug for code path, rm log
* step 8, rename all macro & func from cuda by sycl
* fix error of select non-zero device, format device list
* ren ggml-sycl.hpp -> ggml-sycl.h
* clear CMAKE to rm unused lib and options
* correct queue: rm dtct:get_queue
* add print tensor function to debug
* fix error: wrong result in 658746bb26702e50f2c59c0e4ada8e9da6010481
* summary dpct definition in one header file to replace folder:dpct
* refactor device log
* mv dpct definition from folder dpct to ggml-sycl.h
* update readme, refactor build script
* fix build with sycl
* set nthread=1 when sycl, increase performance
* add run script, comment debug code
* add ls-sycl-device tool
* add ls-sycl-device, rm unused files
* rm rear space
* dos2unix
* Update README_sycl.md
* fix return type
* remove sycl version from include path
* restore rm code to fix hang issue
* add syc and link for sycl readme
* rm original sycl code before refactor
* fix code err
* add know issue for pvc hang issue
* enable SYCL_F16 support
* align pr4766
* check for sycl blas, better performance
* cleanup 1
* remove extra endif
* add build&run script, clean CMakefile, update guide by review comments
* rename macro to intel hardware
* editor config format
* format fixes
* format fixes
* editor format fix
* Remove unused headers
* skip build sycl tool for other code path
* replace tab by space
* fix blas matmul function
* fix mac build
* restore hip dependency
* fix conflict
* ren as review comments
* mv internal function to .cpp file
* export funciton print_sycl_devices(), mv class dpct definition to source file
* update CI/action for sycl code, fix CI error of repeat/dup
* fix action ID format issue
* rm unused strategy
* enable llama_f16 in ci
* fix conflict
* fix build break on MacOS, due to CI of MacOS depend on external ggml, instead of internal ggml
* fix ci cases for unsupported data type
* revert unrelated changed in cuda cmake
remove useless nommq
fix typo of GGML_USE_CLBLAS_SYCL
* revert hip cmake changes
* fix indent
* add prefix in func name
* revert no mmq
* rm cpu blas duplicate
* fix no_new_line
* fix src1->type==F16 bug.
* pass batch offset for F16 src1
* fix batch error
* fix wrong code
* revert sycl checking in test-sampling
* pass void as arguments of ggml_backend_sycl_print_sycl_devices
* remove extra blank line in test-sampling
* revert setting n_threads in sycl
* implement std::isinf for icpx with fast math.
* Update ci/run.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/sycl/run-llama2.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update examples/sycl/run-llama2.sh
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Update CMakeLists.txt
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* add copyright and MIT license declare
* update the cmd example
---------
Co-authored-by: jianyuzh <jianyu.zhang@intel.com>
Co-authored-by: luoyu-intel <yu.luo@intel.com>
Co-authored-by: Meng, Hengyu <hengyu.meng@intel.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* implemented dynamic temperature sampling from koboldcpp
* removed trailing whitespace
* removed unused temp parameter in llama_sample_entropy
* exposed exponent_val in dynamic temp sampler
* added debug check for printf statements
* use nullptr in llama_sample_softmax call during llama_sample_entropy
this avoids counting the time taken stats twice
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* return earlier if there is only 1 candiate (i.e. max_entropy == 0)
* reformat 't' case in llama_sample_queue
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* check for one or zero candidates case in llama_sample_entropy
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Jared Van Bortel <cebtenzzre@gmail.com>
* kl-divergence: be able to save all logits to a file
* Add ability to compute KL-divergence
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* TruthfulQA: 1st attempt, does not look like it is working
The same implementation can be used for HellaSwag as well,
so I converted a HellaSwag validation dataset to the binary
format used here and tested with that. The score is only
around 50, so something is not quite right.
* TruthfulQA: works but the result is bad
I know it works because if I convert the HellaSwag validation
data to the binary format used in the truthful_qa_score() function
I get the exact same result as from the hellaswag_score() function.
But I guess, the questions are tricky and the way I have done
the combination of question + answer is very likely not the best.
The TruthfulQA validation dataset contains 817 questions, with
random chance result around 19%. With this version I get
29.1% for Mistral-7B and 55.2% for Mistral-7B-Instruct-v0.2.
The HF leader board results for these two models are
42.2% and 68.3%, respectively.
* TruthfulQA: fix random sample
* TruthfulQA: prepare tasks in parallel for large test datasets
* Rename truthful_qa to multiple_choice
* Make MSVC happy
I had forgotten that MSVC does not make constexpr's available
inside a lambda.
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* winogrande: simple implementation
It doesn't look like it is working - why?
For Mistral-7B it is barely better than
random chance (score ~60% for 1267 tasks), while I see
Mistral-7B scoring 78.4% on the HF leader board.
1-sigma statistical uncertainty for 1267 tasks is ~1.4,
so no way the difference is due to statistics.
* winogrande: somewhat better
Score for Mistrali7-B is now 68.9 on the validation set of
winogrande_debiased. Still far from the reported 78.4, but
better than what I had before.
* winogrande: improving
Mistral-7B score is now 73.56.
Still not quite 78.4 but getting there.
We are also getting a lower score on HellaSwag
compared to HF leader board, so I'm not expecting
we will get up to 78.4 anyway.
It looks like it is better to skip the choice word(s)
when evaluating the average log-likelihood. This kind of
makes sense because a more common word (in Winogrande this is
often a name) will have a higher probability without knowing
about the follow up context, and this will skew the log-likelihood
towards the more common word. We can only do this if the
choice words are not last in the sentence.
It also looks like it is better to skip the punctuation at the
end of the sentence, provided the choice words are not last.
* winogrande: add dataset instructions
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
* add the parameter : --no-display-prompt , combine with --log-disable it will display only the generated tokens
* remove empty line
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* llama : ggml-backend integration
* ggml-backend : add names to buffers
* fix unmap after loading
* batched-bench : add tensor_split param
* llama : check for null tensor_split
* ggml-backend : increase GGML_MAX_BACKENDS
* improve graph splitting, partial fix for --no-kv-offload
* cuda : add ggml-backend split buffer support
* cuda : do not create buffer types for devices that don't exist (fixes usage without CUDA devices available)
* ggml : fix null backend dereference (#4807)
* ggml : fix null backend dereference
* ggml : also check ggml_backend_is_cpu
* test-backend-ops : check buffer allocation failures
* llama : add cparam (split_mode) and command line argument (--split-mode, -sm) to configure the split mode (none, layer or row)
* ggml : fix mul_mat_id work size
* llama : rewrite session kv load/set without graphs
* minor
* llama : only initialize used backends, free backends on context free
* llama : abort ctx if cuda backend init fails
* llama : rewrite lora with ggml-backend and compute on CPU
ggml-ci
* llama : only map to a backend buffer the region of the file mapping containing the tensors used in the buffer
* opencl : add ggml-backend buffer type
* cuda : only use batched_cublas with batched mat muls (fixes fp16 tg perf)
* llama : on Metal, by default offload the full model
ggml-ci
* metal : page align the data ptr (#4854)
* Apply suggestions from code review
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* cuda : fix split buffer free
* address review comments
* llama-bench : add split-mode parameter
* fix whitespace
* opencl : fix double initialization
* server : add --split-mode parameter
* use async copy and compute to improve multi-gpu performance
ggml-ci
* use async memcpys to copy the graph outputs to the CPU
* fix opencl
* use a host buffer for the cpu compute buffer for faster copies to the gpu
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* common : streamline the formatting of help
- Separate alternative parameters by a comma
- Do not indent `--version` differently
* Update common/common.cpp
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* feat: add avx_vnni based on intel documents
* ggml: add avx vnni based on intel document
* llama: add avx vnni information display
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* docs: add more details about using oneMKL and oneAPI for intel processors
* Update ggml.c
Fix indentation upgate
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Check the full vocab for grammar only if necessary
* Fix missing logit restoration step (?)
Does this matter, actually?
* Fix whitespace / formatting
* Adjust comment
* Didn't mean to push test gbnf
* Split sampling into the helper function (?)
And also revert the changes made to the header
* common : fix final newline
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* initial commit, going through initializations
* main loop finished, starting to debug
* BUG: generates gibberish/repeating tokens after a while
* kv_cache management
* Added colors to distinguish drafted tokens (--color). Updated README
* lookup : fix token positions in the draft batch
* lookup : use n_draft from CLI params
* lookup : final touches
---------
Co-authored-by: Leon Ericsson <leon.ericsson@icloud.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* feat: Allow overriding GGUF metadata when loading model
* Fix the one time GCC is stricter than clang about something
* Step1
* Refactor... basically everything!
* Nuke obsolete GetArrayLen struct
* simplify std::string specialization
* Various cleanups
Add informational output when overrides are applied
Warn user when an override with the wrong type is specified
* Fix broken logic for parsing bool KV overrides
Fix issue where overrides didn't apply when key missing in GGUF metadata
Resolve merge changes
* llama : rearrange model params
* Update new GET_KEY call
Add note that metadata KV overrides aren't reflected in initial metadata KV info dump
---------
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Samplers sequence order w parameter
* Cleaned commented code
* Fixed formatting
* Rewrote with unordered_map
* Revert and rewrite, too many problems and safeguards would be needed
* Fixed code style
* Code style fixes according to review
* More readable samplers input string, fixed help
* Style fix in sampler_queue
* Formatting fixes
* Fixing whitespaces
* cmake : fix joining of REAL_GIT_DIR
* fix includes with help from include-what-you-use
* make : remove unneeded deps and add test-rope target
* fix C includes in C++ source files
* Revert "fix includes with help from include-what-you-use"
This reverts commit 635e9fadfd.
* llama : keep track of used KV cells + better KV cache management
* llama : zero KV cache used upon clear
ggml-ci
* llama : allow exporting a view of the KV cache (#4180)
* Allow exporting a view of the KV cache
* Allow dumping the sequences per cell in common
* Track max contiguous cells value and position as well
* Fix max contiguous empty cells index calculation
Make dump functions deal with lengths or sequences counts > 10 better
* Fix off by one error in dump_kv_cache_view
* Add doc comments for KV cache view functions
Eliminate cell sequence struct; use llama_seq_id directly
Minor cleanups
* common : add -dkvc arg for enabling kv cache dumps
---------
Co-authored-by: Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com>
- introduces help entry for the argument
- cuts '--gpu-layers' form in order to simplify usage and documentation.
Signed-off-by: Jiri Podivin <jpodivin@gmail.com>
Co-authored-by: Jiri Podivin <jpodivin@redhat.com>
* gguf-py: gguf-dump: Respect --no-tensor flag in JSON mode.
* Respect add_bos_token GGUF metadata value
* gguf-py: Try to fix SpecialVocab giving up too easily for the Nth time
* wip llava python bindings compatibility
* add external llava API
* add base64 in-prompt image support
* wip refactor image loading
* refactor image load out of llava init
* cleanup
* further cleanup; move llava-cli into its own file and rename
* move base64.hpp into common/
* collapse clip and llava libraries
* move llava into its own subdir
* wip
* fix bug where base64 string was not removed from the prompt
* get libllava to output in the right place
* expose llava methods in libllama.dylib
* cleanup memory usage around clip_image_*
* cleanup and refactor *again*
* update headerdoc
* build with cmake, not tested (WIP)
* Editorconfig
* Editorconfig
* Build with make
* Build with make
* Fix cyclical depts on Windows
* attempt to fix build on Windows
* attempt to fix build on Windows
* Upd TODOs
* attempt to fix build on Windows+CUDA
* Revert changes in cmake
* Fix according to review comments
* Support building as a shared library
* address review comments
---------
Co-authored-by: M. Yusuf Sarıgöz <yusufsarigoz@gmail.com>
Co-authored-by: Jared Van Bortel <jared@nomic.ai>
* cmake : fix build when .git does not exist
* cmake : simplify BUILD_INFO target
* cmake : add missing dependencies on BUILD_INFO
* build : link against build info instead of compiling against it
* zig : make build info a .cpp source instead of a header
Co-authored-by: Matheus C. França <matheus-catarino@hotmail.com>
* cmake : revert change to CMP0115
---------
Co-authored-by: Matheus C. França <matheus-catarino@hotmail.com>
* Add '-ngl' support to finetune.cpp
* Add fprintf in ggml_cuda_op_add
When I tried CUDA offloading during finetuning following the readme, I got an assert here.
This probably isn't an important case because inference later gives a warning saying you should use f16 or f32 instead when using lora
* Add 'finetune.sh', which currently fails when using GPU
"error: operator (): Finetuning on tensors with type 'f16' is not yet supported"
* tweak finetune.sh
* Suppress some warnings in ggml.c
* Add f16 implementation to ggml_compute_forward_add_f16_f32
* Add an f16 case to ggml_add_cast_impl and llama_build_lora_finetune_graphs
* finetune.sh: Edit comments
* Add "add_f16_f32_f32_cuda"
* Tweak an error message
* finetune.sh: Add an optional LLAMA_MODEL_DIR variable
* finetune.sh: Add an optional LLAMA_TRAINING_DIR variable
* train : minor
* tabs to spaces
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
* Introduce the new Min-P sampler by @kalomaze
The Min-P sampling method was designed as an alternative to Top-P, and aims to ensure a balance of quality and variety. The parameter *p* represents the minimum probability for a token to be considered, relative to the probability of the most likely token.
* Min-P enabled and set to 0.05 default
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: cebtenzzre <cebtenzzre@gmail.com>
* Extend llama_kv_cache_seq_rm to allow matichng any sequence
* Replace llama_kv_cache_tokens_rm with llama_kv_cache_clear
Use llama_kv_cache_clear for cache clearing
Change calls to llama_kv_cache_tokens_rm that want to delete by position to use llama_kv_cache_seq_rm functionality
* added `llama_model_token_*` variants to all the `llama_token_*` functions.
* added `LLAMA_API`
* formatting
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* removed old `llama_token` functions
* changed 3 more functions to take in model
- `llama_token_get_text`
- `llama_token_get_score`
- `llama_token_get_type`
* added back docs
* fixed main.cpp
* changed token functions to use new model variants
* changed token functions to use new model variants
---------
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
