10.7B model, depth upscaled version of two mistral based finetunes

27 Pulls Updated 8 months ago

5eacd9a5ff26 · 137B
{ "num_ctx": 8092, "stop": [ "<|im_end|>", "<|end_of_turn|>", "</s>", "<|im_start|>" ], "temperature": 0.6 }