LLM EXPLORER 60,702 MODELS INDEXED

DeepSeek V4 Flash W4A16 FP8 by pastapaul

By pastapaul · 5589 downloads
HF Repo ↗ Copy URL

DeepSeek V4 Flash W4A16 FP8 is an open-source language model by pastapaul. Features: 44.1b LLM, VRAM: 152.5GB, Context: 1024K, License: apache-2.0, LLM Explorer Score: 0.29.

Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-...   Compressed-tensors   Deepseek   Deepseek-v4   Deepseek v4   En   Fp8-block   Gptq   Mixture-of-experts   Moe   Region:us   Safetensors   Sharded   Tensorflow   Vllm   W4a16   Zh
Model Card on HF 🤗: https://huggingface.co/pastapaul/DeepSeek-V4-Flash-W4A16-FP8 

DeepSeek V4 Flash W4A16 FP8 Parameters and Internals

LLM NameDeepSeek V4 Flash W4A16 FP8
Repository 🤗https://huggingface.co/pastapaul/DeepSeek-V4-Flash-W4A16-FP8 
Base Model(s)  DeepSeek V4 Flash   deepseek-ai/DeepSeek-V4-Flash
Model Size44.1b
Required VRAM152.5 GB
Updated2026-05-17
Maintainerpastapaul
Model Typedeepseek_v4
Model Files  50.0 GB: 1-of-4   50.0 GB: 2-of-4   50.0 GB: 3-of-4   2.5 GB: 4-of-4
Supported Languagesen zh
Model ArchitectureDeepseekV4ForCausalLM
Licenseapache-2.0
Context Length1048576
Model Max Length1048576
Transformers Version5.8.0.dev0
Tokenizer ClassTokenizersBackend
Padding Token<｜end▁of▁sentence｜>
Vocabulary Size129280
Torch Data Typebfloat16

Best Alternatives to DeepSeek V4 Flash W4A16 FP8

Best Alternatives
Context / RAM
Downloads
Likes
DeepSeek V4 Flash W4A16 FP81024K / 152.5 GB262117