mirror of
https://github.com/NVIDIA/Model-Optimizer.git
synced 2026-11-04 08:27:53 +08:00
Chenhan D. YuandClaude Opus 4.6
3131195241
DFlash (Block Diffusion for Flash Speculative Decoding) predicts an entire block of tokens in a single forward pass using masked parallel prediction with KV injection from the target model's hidden states. Key features: - Feature fusion (multi-layer hidden states -> FC + RMSNorm) - KV injection (fused features as K/V in every draft layer with QK-norm) - Random anchor sampling with bidirectional intra-block attention - Logit distillation with exponential loss decay (gamma weighting) - Multi-node DDP training with checkpoint resume - Export to z-lab compatible HF format - Online validation (context-dependent ground truth) Training recipe: modelopt_recipes/general/speculative_decoding/dflash.yaml Results: examples/speculative_decoding/doc/dflash_results.md ### ModelOpt Eval (online validation, osl=512) | Dataset | z-lab | ModelOpt (306K) | Diff | |---------|-------|-----------------|------| | gsm8k | 4.10 | **5.19** | **+1.09** | | MT-Bench | 3.58 | **4.36** | **+0.78** | ### z-lab Official Eval (dflash.benchmark, osl=512) | Dataset | z-lab | ModelOpt (306K) | Diff | |---------|-------|-----------------|------| | gsm8k | **5.00** | 4.08 | -0.92 | | MT-Bench | **3.28** | 2.99 | -0.29 | > z-lab model trained with block_size=16. ModelOpt trained with block_size=8. ## Evaluation Method Impact (gsm8k) | Eval Method | z-lab checkpoint | ModelOpt (306K) | |-------------|-----------------|-----------------| | Fixed GT (ModelOpt eval) | 2.95 | 4.23 | | Online GT (ModelOpt eval) | 4.10 | **5.19** | | z-lab official eval | **5.00** | 4.08 | ### What does this PR do? Type of change: ? <!-- Use one of the following: Bug fix, new feature, new example, new tests, documentation. --> <!-- Details about the change. --> ### Usage ```python # Add a code snippet demonstrating how to use this ``` ### Testing <!-- Mention how have you tested your change if applicable. --> ### Before your PR is "*Ready for review*" Make sure you read and follow [Contributor guidelines](https://github.com/NVIDIA/Model-Optimizer/blob/main/CONTRIBUTING.md) and your commits are signed (`git commit -s -S`). Make sure you read and follow the [Security Best Practices](https://github.com/NVIDIA/Model-Optimizer/blob/main/SECURITY.md#security-coding-practices-for-contributors) (e.g. avoiding hardcoded `trust_remote_code=True`, `torch.load(..., weights_only=False)`, `pickle`, etc.). - Is this change backward compatible?: ✅ / ❌ / N/A <!--- If ❌, explain why. --> - If you copied code from any other sources or added a new PIP dependency, did you follow guidance in `CONTRIBUTING.md`: ✅ / ❌ / N/A <!--- Mandatory --> - Did you write any new necessary tests?: ✅ / ❌ / N/A <!--- Mandatory for new features or examples. --> - Did you update [Changelog](https://github.com/NVIDIA/Model-Optimizer/blob/main/CHANGELOG.rst)?: ✅ / ❌ / N/A <!--- Only for new features, API changes, critical bug fixes or backward incompatible changes. --> ### Additional Information <!-- E.g. related issue. --> <!-- This is an auto-generated comment: release notes by coderabbit.ai --> ## Summary by CodeRabbit * **New Features** * Added DFlash speculative decoding mode with parallel block prediction support. * Included training launchers and MT-Bench evaluation scripts for DFlash models. * Added online acceptance rate validation for improved inference verification. * **Documentation** * DFlash quick start guide with configuration parameters and training examples. * Performance results and benchmarks for DFlash-trained models. <!-- end of auto-generated comment: release notes by coderabbit.ai --> --------- Signed-off-by: Chenhan Yu <chenhany@nvidia.com> Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
56 lines
2.0 KiB
Python
56 lines
2.0 KiB
Python
# SPDX-FileCopyrightText: Copyright (c) 2024 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
#
|
|
# Licensed under the Apache License, Version 2.0 (the "License");
|
|
# you may not use this file except in compliance with the License.
|
|
# You may obtain a copy of the License at
|
|
#
|
|
# http://www.apache.org/licenses/LICENSE-2.0
|
|
#
|
|
# Unless required by applicable law or agreed to in writing, software
|
|
# distributed under the License is distributed on an "AS IS" BASIS,
|
|
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
# See the License for the specific language governing permissions and
|
|
# limitations under the License.
|
|
|
|
import json
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
|
|
@pytest.fixture(scope="session")
|
|
def tiny_conversations_path(tmp_path_factory):
|
|
"""Tiny JSONL with short synthetic conversations for compute_hidden_states_hf tests.
|
|
|
|
Uses minimal single-turn conversations so that tokenized lengths stay well
|
|
within the tiny test model's max_position_embeddings (32) even after chat
|
|
template formatting.
|
|
"""
|
|
tmp_dir = tmp_path_factory.mktemp("tiny_convs")
|
|
output_file = tmp_dir / "train.jsonl"
|
|
conversations = [
|
|
{
|
|
"conversation_id": f"test-{i}",
|
|
"conversations": [
|
|
{"role": "user", "content": "What is 2 plus 2?"},
|
|
{"role": "assistant", "content": "4"},
|
|
],
|
|
}
|
|
for i in range(5)
|
|
]
|
|
with open(output_file, "w") as f:
|
|
f.writelines(json.dumps(conv) + "\n" for conv in conversations)
|
|
return output_file
|
|
|
|
|
|
@pytest.fixture(scope="session", autouse=True)
|
|
def tiny_daring_anteater_path():
|
|
"""Return path to synthetic test data in OpenAI messages format.
|
|
|
|
Uses examples/dataset/synthetic_conversations_1k.jsonl (1000 samples,
|
|
900 single-turn + 100 two-turn). Synthesized by Claude (Anthropic),
|
|
Apache-2.0 licensed.
|
|
"""
|
|
return Path(__file__).parents[3] / "examples" / "dataset" / "synthetic_conversations_1k.jsonl"
|