101 MB

Ctrl+K

1 contributor

History: 9 commits

episod

Card: say that sampling settings affect only the first token with the drafter on (tt-orchard issue 107)

e1ef0ac verified 17 days ago

base_config
Upload folder using huggingface_hub 22 days ago
evidence
Rebuilt with the reviewed graft code; full GSM8K; what the model is for, how to call it, and screenshots of the companion page 21 days ago
images
Card: screenshots retaken with the real model in the current page layout 21 days ago
vllm_models
Upload folder using huggingface_hub 22 days ago
wheels
Upload folder using huggingface_hub 22 days ago
.gitattributes

1.97 kB
Rebuilt with the reviewed graft code; full GSM8K; what the model is for, how to call it, and screenshots of the companion page 21 days ago
README.md

15.3 kB
Card: say that sampling settings affect only the first token with the drafter on (tt-orchard issue 107) 17 days ago
graft.json

920 Bytes
Graft the nearest model's mtp.* head: drafter on, sampling on the device; full IFEval and GSM8K subset, speed rows 21 days ago
install.sh

8.18 kB
Upload folder using huggingface_hub 22 days ago
model.py

421 Bytes
Upload folder using huggingface_hub 22 days ago
prepare_model_dir.py

12 kB
Rebuilt with the reviewed graft code; full GSM8K; what the model is for, how to call it, and screenshots of the companion page 21 days ago
requirements.txt

200 Bytes
Upload folder using huggingface_hub 22 days ago
run.sh

9.9 kB
Graft the nearest model's mtp.* head: drafter on, sampling on the device; full IFEval and GSM8K subset, speed rows 21 days ago
tt_kernel_manifest.json

2.39 kB
Rebuilt with the reviewed graft code; full GSM8K; what the model is for, how to call it, and screenshots of the companion page 21 days ago
vllm-overrides.txt

445 Bytes
Upload folder using huggingface_hub 22 days ago