elyza/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b
Full fine-tune vs llm-jp/llm-jp-4-32b-a3b-base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 129/129, MLP 2741/2741, Attention 128/128, lm_head 1/1
3 derivatives checked
- na0x2c6/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b-GGUF: Same weights, converted
- mmnga-o/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b-gguf: Quantization
- rariruluis/ELYZA-Thinking-1.0-llm-jp-4-32b-a3b-mlx-4bit: Quantization