Qwen/Qwen2-57B-A14B-Instruct
Partial fine-tune vs Qwen/Qwen2-57B-A14B. Partial change: 80% of the weights changed (attention, embeddings, lm_head, mlp, norms), in 28 of 28 layers.
Tensors changed: Embeddings 1/1, MLP 2059/2745, Attention 196/196, Norms 51/57, lm_head 1/1
6 derivatives checked
- legraphista/Qwen2-57B-A14B-Instruct-GGUF: Same weights, converted
- mradermacher/Qwen2-57B-A14B-Instruct-GGUF: Quantization
- mradermacher/Qwen2-57B-A14B-Instruct-i1-GGUF: Quantization
- Qwen/Qwen2-57B-A14B-Instruct-GPTQ-Int4: Quantization
- tensorblock/Qwen2-57B-A14B-Instruct-GGUF: Quantization
- legraphista/Qwen2-57B-A14B-Instruct-IMat-GGUF: Same weights, converted