LLM EXPLORER 65,548 MODELS INDEXED

DSV4.1 Flash 1017 DERISKED EXL3 2.50bpw by vcruz305

By vcruz305 · 63 downloads
HF Repo ↗ Copy URL

DSV4.1 Flash 1017 DERISKED EXL3 2.50bpw is an open-source language model by vcruz305. Features: 550.9b LLM, VRAM: 387.7GB, License: mit, LLM Explorer Score: 0.3.

Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-...   Conversational   Deepseek   Deepseek v41   Exl3   Exllamav3   Fp8   Image-text-to-text   Mixed-precision   Moe   Quantization   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/vcruz305/DSV4.1-Flash-1017-DERISKED-EXL3-2.50bpw 

DSV4.1 Flash 1017 DERISKED EXL3 2.50bpw Parameters and Internals

LLM NameDSV4.1 Flash 1017 DERISKED EXL3 2.50bpw
Repository 🤗https://huggingface.co/vcruz305/DSV4.1-Flash-1017-DERISKED-EXL3-2.50bpw 
Base Model(s)  deepseek-ai/DeepSeek-V4.1-Flash-1017   deepseek-ai/DeepSeek-V4.1-Flash-1017
Model Size550.9b
Required VRAM387.7 GB
Updated2026-10-25
Maintainervcruz305
Model Typedeepseek_v41
Model Files  8.5 GB: 1-of-17   8.5 GB: 2-of-17   8.5 GB: 3-of-17   8.5 GB: 4-of-17   8.5 GB: 5-of-17   8.5 GB: 6-of-17   8.5 GB: 7-of-17   8.5 GB: 8-of-17   8.5 GB: 9-of-17   8.5 GB: 10-of-17   8.5 GB: 11-of-17   8.5 GB: 12-of-17   8.5 GB: 13-of-17   8.5 GB: 14-of-17   64.7 GB: 15-of-17   102.0 GB: 16-of-17   102.0 GB: 17-of-17
Model ArchitectureDeepseekV41ForCausalLM
Licensemit
Model Max Length1048576
Transformers Version5.6.0
Tokenizer ClassPreTrainedTokenizerFast
Beginning of Sentence Token<｜begin▁of▁sentence｜>
End of Sentence Token<｜end▁of▁sentence｜>