reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B

Partial fine-tune vs Qwen/Qwen3-1.7B. Partial change: 85% of the weights changed (attention, lm_head, mlp, norms), in 28 of 28 layers.

Tensors changed: Embeddings 0/1, Norms 85/113, MLP 84/84, Attention 112/112, lm_head 1/1

3 derivatives checked

• reaperdoesntknow/DistilQwen3-1.7B-uncensored: Partial fine-tune
• tinyopsec/Qwen3-1.7B-Distilled-30B-A3B-GGUF: Same weights, converted
• reaperdoesntknow/Qwen3-1.7B-Distilled-30B-A3B-SFT: Partial fine-tune

On the Hub · Downloads on Model Pulse