ehristoforu/mistral-distil-test-0

Partial fine-tune (fewer layers) vs unsloth/Mistral-Nemo-Instruct-2407. ehristoforu/mistral-distil-test-0 matches part of unsloth/Mistral-Nemo-Instruct-2407 (228 of its 363 tensors; the rest of unsloth/Mistral-Nemo-Instruct-2407 has no counterpart). On the shared part: Partial change: 60% of the weights changed (attention, mlp, norms), in 18 of 25 layers.

Tensors changed: Embeddings 0/1, Norms 36/51, MLP 54/75, Attention 72/100, lm_head 0/1

On the Hub · Downloads on Model Pulse