The aim of this model is to retain the reasoning capabilities of DeepSeek-R1-Distill-Llama-70B, while aligning more with the original Llama 3.3 model on which it is based.
As this model derives from Llama 3.3, the Llama 3.3 Community License Agreement applies.
Use the DeepSeek Chat Prompt Template when working with this model.
70B Safetensors BF16 format:
Use with transformers as you would Llama 3.3, but use the DeepSeek Chat Prompt Template as you would with the original DeepSeek-R1-Distill-Llama models.
70B GGUF Quantised versions:
Use these with Llama.cpp, LM Studio or Kobold.cpp. Thanks to mradermacher for converting these from the safetensors format.
| Filename | Type | Size | Quality |
|---|---|---|---|
| LlamaAligned-DeepSeekR1-Distill-70b-Q4_K_M.gguf | Q4_K_M | 42.5GB | OK quality, default. |
| LlamaAligned-DeepSeekR1-Distill-70b-Q8_0.gguf part1 part2 | Q8_0 | 75.0GB | Best quality quantised version. |
| LlamaAligned-DeepSeekR1-Distill-70b-Q3_K_S.gguf | Q3_K_S | 30.9GB | Lower quality. |
8B Safetensors BF16 format:
Use with transformers as you would Llama 3.1, but use the DeepSeek Chat Prompt Template as you would with the original DeepSeek-R1-Distill-Llama models.
Use model id BlueBeck/LlamaAligned-DeepSeekR1-Distill-8b
8B GGUF Quantised versions:
Use these with Llama.cpp, LM Studio or Kobold.cpp. Thanks to mradermacher for converting these from the safetensors format.
| Filename | Type | Size | Quality |
|---|---|---|---|
| LlamaAligned-DeepSeekR1-Distill-8b-Q4_K_M.gguf | Q4_K_M | 4.92GB | OK quality, default. |
| LlamaAligned-DeepSeekR1-Distill-8b-Q8_0.gguf | Q8_0 | 8.54GB | Best quality quantised version. |
| LlamaAligned-DeepSeekR1-Distill-8b-Q6_K.gguf | Q6_K | 6.6GB | High quality. |
| LlamaAligned-DeepSeekR1-Distill-8b-Q5_K_M>.gguf | Q5_K_M> | 5.73GB | Good quality. |
| LlamaAligned-DeepSeekR1-Distill-8b-Q3_K_S.gguf | Q3_K_S | 3.66GB | Lower quality. |
- Downloads last month
- 23
Model tree for BlueBeck/LlamaAligned-DeepSeekR1-Distill-70b
Base model
deepseek-ai/DeepSeek-R1-Distill-Llama-70B