You can deploy this model and start earning money today!

inference-optimization/
ctest-Qwen3-8B-speculator.eagle3-fp8ablation-bf16

Name: ctest-Qwen3-8B-speculator.eagle3-fp8ablation-bf16
Availability: InStock

Views

No views yet

ctest: Qwen3-8B EAGLE-3 speculator — FP8 hidden-states ablation (bf16)

Test checkpoint — not a production release. Trained to validate the FP8-quantized hidden-states transfer backend added in vllm-project/speculators#1028 (draft, reviving #491).

Verifier: Qwen/Qwen3-8B
Algorithm: eagle3
Hidden-states precision during training: bf16 (existing bf16 file backend)
Training data: 5K magpie + 5K ultrachat samples from inference-optimization/Dataset-Qwen3-235B-Instruct

This is one of a matched pair (...-fp8ablation-bf16 / ...-fp8ablation-fp8) trained identically except for the hidden-states transfer precision, to isolate the effect of FP8 quantization on speculator quality. See the sibling repo for the other precision.

Validation metrics (final epoch)

loss_0_epoch: 1.059357
full_acc_0_epoch: 0.694162
cond_acc_0_epoch: 0.694162
loss_1_epoch: 1.374979
full_acc_1_epoch: 0.470217
cond_acc_1_epoch: 0.677388
loss_2_epoch: 1.575933
full_acc_2_epoch: 0.315562
cond_acc_2_epoch: 0.671088
loss_epoch: 4.010267

guidellm serving eval

See acceptance.csv in this repo for the full per-subset guidellm breakdown (9 subset rows).

Full ablation writeup, all 6 checkpoints' results, exact commands, and code: see the results package referenced from PR #1028.

inference-optimization/ctest-Qwen3-8B-speculator.eagle3-fp8ablation-bf16

ctest: Qwen3-8B EAGLE-3 speculator — FP8 hidden-states ablation (bf16)

Validation metrics (final epoch)

guidellm serving eval

inference-optimization/
ctest-Qwen3-8B-speculator.eagle3-fp8ablation-bf16