This will be necessary to support Jamba
(and other recurrent models mixed with Attention).
Doesn't compile yet, and finding a slot isn't yet done correctly for recurrent states.
* initial commit for sealion support
* add sealion support
* minor fix
* q/k ln and pos_embd only if required
* Apply suggestions from code review
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* minor : clear whitespaces
---------
Co-authored-by: bryan <bryansiow@aisingapore.org>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* Add openchat chat template
* Add chat template test for openchat
* Add chat template for vicuna
* Add chat template for orca-vicuna
* Add EOS for vicuna templates
* Combine vicuna chat templates
* Add tests for openchat and vicuna chat templates
* Add chat template for alpaca
* Add separate template name for vicuna-orca
* Remove alpaca, match deepseek with jinja output
* Regenerate chat template test with add_generation_prompt
* Separate deepseek bos from system message
* Match openchat template with jinja output
* Remove BOS token from templates, unprefix openchat
* 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>
* Fix Vulkan no kv offload incoherence
* Add k-quant mul mat mat shaders
* Rework working buffer allocation, reduces vram use noticeably
Clean up cpu assist code, replaced with ggml-backend offload function
* Default to all dedicated GPUs
* Add fallback for integrated GPUs if no dedicated GPUs are found
* Add debug info which device is allocating memory
* Fix Intel dequant issue
Fix validation issue
* Fix Vulkan GGML_OP_GET_ROWS implementation
* Clean up merge artifacts
* Remove Vulkan warning
* Support xverse model convert to gguf format.
* 1. Convert xverse models to gguf;
2. Add LLM_ARCH_XVERSE inference in llama.cpp;
3. Add xverse item in Supported models in README.md;
* * gguf-py: remove redundant logs
* llama: remove the init_mapping_prefetch custom parameter
* llama.cpp: Include the changes from #6122 to exclude the unused outputs of the last layers.
* - Fix format issues
- Remove duplicate set kqv_out to llm_build_kv
* Update llama.cpp
---------
Co-authored-by: willhe <willhe@xverse.cn>
Co-authored-by: willhe <hexin@xverse.cn>
* llama: remove redundant reshape in build_kv_store
This commit removes the reshape of the V matrix in the build_kv_store.
The motivation for this is that V matrix has the shape:
```console
(gdb) p *v_cur
$46 = {type = GGML_TYPE_F32, backend = GGML_BACKEND_TYPE_CPU,
buffer = 0x0, ne = {4096, 512, 1, 1}, nb = {4, 16384, 8388608,
8388608}, op = GGML_OP_MUL_MAT, op_params = {
0 <repeats 16 times>}, flags = 0, grad = 0x0,
src = {0xb496b0, 0x7ffef1c40950, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0, 0x0}, perf_runs = 0, perf_cycles = 0, perf_time_us = 0,
view_src = 0x0, view_offs = 0, data = 0x0,
name = "Vcur-0", '\000' <repeats 57 times>, extra = 0x0,
padding = "\000\000\000\000\000\000\000"}
```
And after reshaping this tensor we get:
```console
gdb) p *ggml_reshape_2d(ctx, v_cur, n_embd_v_gqa, n_tokens)
$44 = {type = GGML_TYPE_F32, backend = GGML_BACKEND_TYPE_CPU,
buffer = 0x0, ne = {4096, 512, 1, 1}, nb = {4, 16384, 8388608,
8388608}, op = GGML_OP_RESHAPE, op_params = {
0 <repeats 16 times>}, flags = 0, grad = 0x0,
src = {0x7ffef1c40e00, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0x0}, perf_runs = 0, perf_cycles = 0, perf_time_us = 0,
view_src = 0x7ffef1c40e00, view_offs = 0, data = 0x0,
name = "Vcur-0 (reshaped)", '\000' <repeats 46 times>, extra = 0x0,
padding = "\000\000\000\000\000\000\000"}
```
I noticed that the `src` and `view_src` fields are different but that the
dimensions are the same. From the code comment it seems like the reshape
call is not needed and perhaps the above can motivate the removal of the
reshape call.
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
* llama : add assert
---------
Signed-off-by: Daniel Bevenius <daniel.bevenius@gmail.com>
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>
* 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>
* backend : offload large batches to GPU
* fix hip
* code cleanup
* fix CUDA split buffers
* Update ggml-backend-impl.h
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* cuda : fix memset without set_device
* imatrix : remove sched affix from weight names
* sched : add a new split if the current one has too many inputs
reduce max inputs per split
more cleanup
* update backends
ggml-ci
---------
Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
* 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>
Information about the Command-R 35B model (128k context) can be found at:
https://huggingface.co/CohereForAI/c4ai-command-r-v01
Based on the llama2 model with a few changes:
1) New hyper parameter to scale output logits (logit_scale)
2) Uses LayerNorm instead of RMSNorm
3) Transfomer layers have a single shared LayerNorm that feeds into both the
self-attention and FFN layers in parallel. There is no post-attention LayerNorm.
4) No support for Rotary Position Embeddings (RoPE) scaling
5) No biases used
Find GGUF files here:
https://huggingface.co/andrewcanis/c4ai-command-r-v01-GGUF
To convert model to GGUF format yourself:
1) Download Command-R Hugging Face safetensors:
git lfs install
git clone https://huggingface.co/CohereForAI/c4ai-command-r-v01
2) Run:
python3 convert-hf-to-gguf.py --outtype f16 ./c4ai-command-r-v01
* additional methods to read model and ctx parameters
* vocab size as a part of a model metadata
* models without vocabulary, convert.py part
* models without vocabulary, llama.cpp part
* PR clean up
* converter scrypt fixes
* llama_vocab_type update (renamed the new key)
* pr review fixes
* revert function renaming
* one more NoVocab assert
* attempt to reduce the impact of a worst-case scenario
* fragmentation calculation fix
* Update llama.cpp
---------
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>
* windows arm ci
* fix `error C2078: too many initializers` with ggml_vld1q_u32 macro for MSVC ARM64
* fix `warning C4146: unary minus operator applied to unsigned type, result still unsigned`
* fix `error C2065: '__fp16': undeclared identifier`
* add gritlm example
* gritlm results match
* tabs to spaces
* comment out debug printing
* rebase to new embed
* gritlm embeddings are back babeee
* add to gitignore
* allow to toggle embedding mode
* Clean-up GritLM sample code.
* Fix types.
* Flush stdout and output ending newline if streaming.
* mostly style fixes; correct KQ_mask comment
* add causal_attn flag to llama_cparams
* gritml : minor
* llama : minor
---------
Co-authored-by: Douglas Hanley <thesecretaryofwar@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
* perplexity : support using multiple sequences to allow larger batch sizes
ggml-ci
* set cparams.n_parallel to the number of sequences
* print tested n_ctx, add assert
* 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`.
This is to support model configurations with "tie_word_embeddings" set to true.
Co-authored-by: Don Mahurin <2797413+dmahurin@users.noreply.github.com>
* server : refactoring (wip)
* server : remove llava/clip objects from build
* server : fix empty prompt handling + all slots idle logic
* server : normalize id vars
* server : code style
* server : simplify model chat template validation
* server : code style
* server : minor
* llama : llama_chat_apply_template support null buf
* server : do not process embedding requests when disabled
* server : reorganize structs and enums + naming fixes
* server : merge oai.hpp in utils.hpp
* server : refactor system prompt update at start
* server : disable cached prompts with self-extend
* server : do not process more than n_batch tokens per iter
* server: tests: embeddings use a real embeddings model (#5908)
* server, tests : bump batch to fit 1 embedding prompt
* server: tests: embeddings fix build type Debug is randomly failing (#5911)
* server: tests: embeddings, use different KV Cache size
* server: tests: embeddings, fixed prompt do not exceed n_batch, increase embedding timeout, reduce number of concurrent embeddings
* server: tests: embeddings, no need to wait for server idle as it can timout
* server: refactor: clean up http code (#5912)
* server : avoid n_available var
ggml-ci
* server: refactor: better http codes
* server : simplify json parsing + add comment about t_last
* server : rename server structs
* server : allow to override FQDN in tests
ggml-ci
* server : add comments
---------
Co-authored-by: Pierrick Hymbert <pierrick.hymbert@gmail.com>
