---
tags:
- qwen3
- eagle3
- speculative-decoding
- draft-model
base_model:
- Qwen/Qwen3-8B
- AngelSlim/Qwen3-8B_eagle3
license: apache-2.0
---

# AdaSPEC — EAGLE3 Draft Model for Qwen3-8B

Reimplementation of AdaSPEC (adaptive speculative decoding) training objective. Baseline comparison.

Part of a course project evaluating per-step weighted loss functions for training
EAGLE3 draft models. Full pipeline and source:
**https://github.com/XLOverflow/anlp_course_project**

Collection: [Qwen3 EAGLE3 — Weighted Loss Variants](https://huggingface.co/collections/XLOverflow/qwen3-eagle3-weighted-loss-variants)

## Training

- **Framework:** [SpecForge](https://github.com/sgl-project/SpecForge) (our fork: https://github.com/XLOverflow/SpecForge)
- **Target model:** `Qwen/Qwen3-8B`
- **Draft init:** `AngelSlim/Qwen3-8B_eagle3`
- **Data:** ShareGPT-style reasoning traces (see `scripts/data/` in project repo)
- AdaSPEC adaptive loss (see paper)
- Initialized from: `baseline-uniform/epoch_4_step_82000`

## Evaluation (Qwen3-8B target)

| Dataset | τ (accept. length) | Speedup | Accuracy |
|---|---|---|---|
| GSM8K | 6.856 | 4.289× | 95.15% |
| MATH500 | 6.678 | 4.206× | 94.40% |

Baselines for reference: Vanilla ≈ 1× speedup, EAGLE-orig ≈ 2× speedup.


## Files

- `model.safetensors` — draft model weights (~763 MB)
- `config.json` — model config
- Corresponds to: `outputs/eagle3-adaspec/epoch_0_step_17026` in the original training output

Optimizer state (~3 GB) is not uploaded — use the project repo's training scripts to resume from scratch if needed.

## Usage

```python
from huggingface_hub import snapshot_download
draft_path = snapshot_download(repo_id="XLOverflow/qwen3-eagle3-adaspec")
# Then load with EAGLE's EaModel — see scripts/eval/eval_combined.py in the project repo.
```