You can deploy this model and start earning money today!

taehyeonkim/
fp16-bf16-wfp16a16kvfp16-wbf16abf16kvbf16

Views

No views yet

FP16/BF16 W-FP16 / A-FP16 / KV-FP16 and W-BF16 / A-BF16 / KV-BF16 (Llama-3.1-8B-Instruct)

This dataset contains FP16 and BF16 reference artifacts for Llama-3.1-8B-Instruct, stored in the same layout as the FP8 W-FP8/A-FP16/KV-FP8 artifact.

w_of_wfp16a16kvfp16_llama_31_8b/ — FP16 weights

The FP16 weights of Llama-3.1-8B-Instruct. Stored per layer: layer_0.safetensors ... layer_31.safetensors + embeddings.safetensors. The 7 linears per layer are stored as FP16… See the full description on the dataset page: https://huggingface.co/datasets/taehyeonkim/fp16-bf16-wfp16a16kvfp16-wbf16abf16kvbf16.