canada-quant
/

DeepSeek-V4-Flash-W4A16-FP8-MTP

Text Generation

compressed-tensors

speculative-decoding

Model card Files Files and versions

DeepSeek-V4-Flash-W4A16-FP8-MTP / tokenizer.json

pastapaul's picture

Phase 2 full GPTQ + BF16 MTP: 89.1% MTP acceptance

e910552 verified 1 day ago

history contribute delete

10.1 MB

File too large to display, you can check the raw version instead.