allenai/tmax-27b

Full fine-tune vs Qwen/Qwen3.6-27B. Fine-tune: 91% of the weights changed, across 64 of 64 layers.

Tensors changed: Embeddings 0/1, MLP 192/192, Attention 355/448, lm_head 0/1, Norms 123/209

1 derivatives checked

• pearsonkyle/tmax-27b-imatrix-MTP-GGUF: Quantization

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse