* Remove vLLM dependency for CUDA
This change adds `attention-kernels` as a dependency for paged
attention and cache reshaping. With that, we don't use vLLM
anywhere for CUDA.
Tested run (since we don't have paged attention in CI):
```
❯ ATTENTION=paged python -m pytest integration-tests -k "llama and awq" --release
[...]
5 snapshots passed.
```
* Fix clippy warning
compressed-tensors is a safetensors extension for sparse, quantized
tensors. The format is more powerful than earlier AWQ/GPTQ/FP8
quantization, because
- Different quantizer configurations can be used for different targets.
- The format can specify input/output quantizers in addition to weight
quantizers.
- Configurable exclusions for quantization.
This change adds a dependency on the `compressed-tensors` package for
its configuration parsing and layer matching functionality.
The following types of quantization are supported in this PR:
- W8A16 and W4A16 INT using GPTQ-Marlin kernels.
- W8A8 and W8A16 FP using FP8-Marlin and cutlass kernels.
Support for other quantization types will be added in subsequent PRs.
fix incorrect output of Qwen2-7B-Instruct-GPTQ-Int4 and Qwen2-7B-Instruct-AWQ
ipex kernel provide func like add_bias, so no need add it outside
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* feat: add support for qwen2 vl model
* feat: fix token padding, enable warmup and process basic request
* fix: improve get_position_ids, add lift embed_tokens
* fix: remove get_cos_sin_hack dev function
* feat: add simple test chat with meesage and text
* fix: lint test
* fix: adjust positional embeddings for multi dimensional position ids
* fix: update docs and lint unused vars
* fix: include linted file
* fix: add norm after text output
* fix: format model file
* fix: adjust for ruff lints
* fix: remove unused rotate_half
* feat: refactors and calc num features
* fix: prefer position_ids passed from vlm causal lm and reset ids on batch
* fix: adjust get_position_ids if not available and add required args to signatures
* fix: adjust resize case for qwen2_vl warmup
* fix: avoid qwen2 vl specific paths with qwen2
* Switch from fbgemm-gpu w8a8 scaled matmul to vLLM/marlin-kernels
Performance and accuracy of these kernels are on par (tested with Llama
70B and 405B). Removes a dependency and resolves some stability issues
we have been seeing.
* Update test snapshots
* Add support for FP8 KV cache scales
Since FP8 only has limited dynamic range, we can scale keys/values
before storing them into the cache (and unscale them in attention). To
avoid rescaling the cache as the absmax values change, good scales are
usually determined per layer using calibration calibration data and stored
in the checkpoint.
This change adds support for for using key-value scales and loading them
from checkpoints in the two most common formats:
- Separate per-layer `k_scale` and `v_scale` scalars.
- Per-layer `kv_scale` scalar (older format).
Currently, scales are only used with an `float8_e4m3fn` cache.
Besides adding support for key/value scales, the `fp8_quantize` function
is also extended to support quantization with a kernel vendored from
vLLM. This is slightly faster than the PyTorch implementation, but also
scales in FP32, potentially improving accuracy.
* Update FP8 KV cache test to use checkpoint with scales
* `can_scale`: check that the attention is flashinfer
Change `fp8_quantize` so that we can pass around reciprocals everywhere,
so scales are always passed around in the checkpoint format.
I also noticed that we ignore any input scales that we might have when
fbgemm is available. Skip this path if we already have a scale.
* add gptq and awq int4 support in intel platform
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix ci failure
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* set kv cache dtype
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* refine the code according to the review command
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* Simplifying conditionals + reverting integration tests values.
* Unused import
* Fix redundant import.
* Revert change after rebase.
* Upgrading the tests (TP>1 fix changes to use different kernels.)
* Update server/text_generation_server/layers/gptq/__init__.py
---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Co-authored-by: Wang, Yi A <yi.a.wang@intel.com>
* Simplify the `attention` function
- Use one definition rather than multiple.
- Add `key`/`value` arguments, so that we don't need the
`PREFILL_IN_KVCACHE` constant.
- Make it kwargs-only (to avoid mixing up the various `Tensor` args).
* Fixup flashinfer support
* Add basic FP8 KV cache support
This change adds rudimentary FP8 KV cache support. The support is
enabled by passing `--kv-cache-dtype fp8_e5m2` to the launcher. Doing so
uses this type for the KV cache. However support is still limited:
* Only the `fp8_e5m2` type is supported.
* The KV cache layout is the same as `float16`/`bfloat16` (HND).
* The FP8 KV cache is only supported for FlashInfer.
* Loading of scales is not yet supported.
* Fix Cargo.toml
* feat: support phi3.5 moe model loading
* fix: prefer llama base model and improve rotary logic
* feat: return reasonable generation and add integration test
* fix: run lint and update docs
* fix: rerun lint for openapi docs
* fix: prefer do_sample false unless temp is set by user, and update chat tests
* fix: small typo adjustments
* fix: consolidate long rope paths
* fix: revert greedy by default and test changes
* Vendor configuration so that we don't have to `trust_remote_code`
* Use SparseMoELayer
* Add support for dense MoE
* Some type annotations
* Add the usual model tests
* Ruff.
---------
Co-authored-by: Daniël de Kok <me@danieldk.eu>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
This change add support for MoE models that use GPTQ quantization.
Currently only models with the following properties are supported:
- No `desc_act` with tensor parallelism, unless `group_size=-1`.
- No asymmetric quantization.
- No AWQ.
* Improve support for GPUs with capability < 8
- For models that cannot use flashinfer, use flash-attn v1 + paged
attention for models with a compute capability older than 8.
- Disable prefix caching when using paged attention.
- When using flash-attn v1, pass the key/value, rather than the
cache, since v1 cannot use block tables.
* nix: add flash-attn-v1 to the server environment
* Move disabling prefix caching into the block of exceptions
* Capability as `usize`s
* Add support for scalar FP8 weight scales
* Support LLM compressor FP8 checkpoints on H100
On H100, we use fbgemm-gpu, which requires bfloat16 as the input dtype.
However, we wouldn't pick up fp8 quantization for models quantized with
LLM compressor. This change adds enough parsing to detect if models have
FP8-quantized weights.
* Remove stray debug print
* Move to moe-kernels package and switch to common MoE layer
This change introduces the new `moe-kernels` package:
- Add `moe-kernels` as a dependency.
- Introduce a `SparseMoELayer` module that can be used by MoE
models.
- Port over Mixtral and Deepseek.
* Make `cargo check` pass
* Update runner
fix regression caused by attention api change. ipex.varlen_attention does not support paged-cache
format kv input now.
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* Tied embeddings in MLP speculator.
* Fixing the scale_weight when users decide to not use the speculation as
much as defined in the config.
* Adding scaling support + optimize some ops.
* Making prefix/flashinfer the default and testing the full release tests.
* Include flashinfer in the docker.
* Using prebuilt.
* Allowing window_left_size (dummy version).
* Disabling flashinfer/prefix caching on odd head_dim
* Disable prefix caching for lora.
* More specific codes.
* Update lock
* Updating integration tests with new values with FI/FD.
Remove paged as a default too, and using FD everywhere.
* Update cargo lock ?
* Upgrade to 1.80 because of bitstream...
* Everywhere 1.80
* Forgot last default place.
* Apply suggestions from code review
Co-authored-by: drbh <david.richard.holtz@gmail.com>
* Updated flake lock
* Tmp
* Upgrade resolution system for less errors in resolution.
* Remove lambda for cleaner function.
* Handling debugger.
* OVerride the env in server tests.
* Is this enough to make it work ?
* This seems to be working.
* Downgrade some logs.
* Fixing the default for vlm.
* Don't enable prefix caching on VLM just yet.
* Change `add_special_tokens` in order to have the correct tokens for chat
input and not (since it's super important with the prefixing now)
* Fixing prefix caching for flashdecoding.
* Update all models.
* Fixed flashinfer version.
* add_special_tokens is internal only
* Fixing seqlen with the new vlms.
* Fixing the issue with `add_special_tokens` not being passed around.
* Fixing the test.
* Removing encoder_decoder (seq2seq).
* Update the chat test.
* Fixing the batching tokenization in flash causal lm.
* Truncating left for radix purposes.
* Oops this doesn't belong here.
* Put back default pure shell.
* Update server tests
- Default to throughput test in k6
- Use TGI_WIGGLE_ROOM to adjust wiggle room
* Only n_heads / process_group.size() are necessary.
* Revert the integrationt tests change (seem linked to head_size
modification).
* Adding error message when assert is violated.
* Fixing the free algorithm to handle times where the common prefix is
smaller.
* Apply suggestions from code review
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
* Update server/text_generation_server/layers/attention/common.py
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
* Fix disabling prefix caching - Fix windowing checks.
* Revert the Cohere tokenizer change (for now using a revision instead).
* Fmt.
---------
Co-authored-by: drbh <david.richard.holtz@gmail.com>
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
This change adds support for FlashInfer. FlashInfer can be enabled using
`FLASH_INFER=1` and is currently only implemented in `FlashCausalLM`.
Since this functionality is currently only for testing, FlashInfer is
not installed anywhere yet.
The FlashInfer API is quite different from FlashAttention/vLLM in that
it requires more global bookkeeping:
* A wrapper class needs to be contstructed (which we just call *state*).
Since this is fairly expensive (due to pinned host memory allocation),
we only do this once in a FlashCausalLM instance or for each CUDA
Graph size.
* Each model forward call needs to be wrapped in `begin_forward` and
`end_forward`. This sets up data structures that can be reused for all
calls to attention for that forward call.
When calling attention, we need access to the state object. To avoid
passing an argument down the call chain (which would require changes to
all models), we use a context variable.
Each model forward call is wrapped using a context manager that does all
the bookkeeping for such a call:
* Set the context variable to the forward call's state.
* Call `begin_forward` on the state.
* Yield.
* Call `end_forward` on the state.
* Reset the context variable.
We cannot use a single shared global variable for this, since e.g. CUDA
Graphs of different sizes each have their own state.
* hotfix: fix xpu crash brought by code refine. torch.xpu rely on import ipex
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* reable gemma2 in xpu
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
* fix in regression in ipex flashattention
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
---------
Signed-off-by: Wang, Yi A <yi.a.wang@intel.com>
Co-authored-by: Wang, Yi A <yi.a.wang@intel.com>
* fix: attempt forward on flash attn2 to check hardware support
* fix: warn window_size_left when using flash attn 1
* fix: prefer version check over test op and avoid window_size_left if not flash attn2
* fix: improve condtional and error message
* fix: update sliding window conditional
* fix: simplify changes and revert model changes
* fix: avoid changing conditional
* fix: typo tweak
- Always return the hidden states.
- Create the output tensor inside the `attention` and `paged_attention`
functions.
This removes the difference between how the output is handled between
attention (output parameter) and paged attention (return value). This
also removes the assumption that the attention implementation can
write to an output tensor (in preparation of FlashInfer).
The `GPTWeightLoader` was structured like this in pseudocode:
if marlin:
Set up tensors in a way that GPTQ-Marlin expects
else:
Set up tensors in a way that ExLlama/GPTQ/AWQ expect
However, the GPT-Marlin implementation details should really be in the
`marlin` module. So move the former part out to a separate
`GPTQMarlinWeightsLoader`.
* Fix GPTQ autotune data type to be compatible with Torch 2.4.0
* Update poetry lock file
* Fix small PaliGemma logprob differences after the torch update
Daniël de Kok