EQUES/TinyDeepSeek-JP-1.5B

Different architecture vs cyberagent/DeepSeek-R1-Distill-Qwen-14B-Japanese. Different architecture: only 0 of 579/338 tensors match by name and shape.

1 derivatives checked

• mradermacher/TinyDeepSeek-JP-1.5B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse