DeepSeek
Coder V2 Lite 16B-A2.4B
codeMoE
15,7B
Parameters · 2,4B active
9,7 Go
Q4_K_M file
≥ 11,2 Go
Recommended VRAM
163 840
Max context
Not measured yet
No card has run the TokenBench protocol on this model yet. The current protocol covers a selection of models; benchmarking any catalog model from the app is coming soon.
Measure my card →Which cards should it fit on?
Cards on the site whose video memory covers the recommended 11,2 Go (file + 4,000-token context). An estimate, not a measurement. Click a card for the detailed estimate.
Technical sheet
- Published
- 2024-06-14
- Architecture
- 27 layers · 16 KV heads · dimension 128
- Attention
- Latent (MLA): the context cache is compressed — much lighter than the standard formula.
- GGUF file
- bartowski/DeepSeek-Coder-V2-Lite-Instruct-GGUF · DeepSeek-Coder-V2-Lite-Instruct-Q4_K_M.gguf Download (9,7 Go)
- Official repo
- deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
Public Hugging Face data (official config.json, GGUF repo). The context cache only counts full-attention layers.