moonshotai/Kimi-Dev-72B
Full fine-tune vs Qwen/Qwen2.5-72B. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Embeddings 1/1, MLP 240/240, Attention 560/560, Norms 161/161, lm_head 1/1
18 derivatives checked
- unsloth/Kimi-Dev-72B-GGUF: Same weights, converted
- bartowski/moonshotai_Kimi-Dev-72B-GGUF: Quantization
- mlx-community/Kimi-Dev-72B-4bit-DWQ: Quantization
- mradermacher/Kimi-Dev-72B-i1-GGUF: Quantization
- mradermacher/Kimi-Dev-72B-GGUF: Quantization
- gabriellarson/Kimi-Dev-72B-GGUF: Quantization
- bullerwins/Kimi-Dev-72B-GGUF: Quantization
- DevQuasar/moonshotai.Kimi-Dev-72B-GGUF: Same weights, converted
- featherless-ai-quants/moonshotai-Kimi-Dev-72B-GGUF: Quantization
- QuantTrio/Kimi-Dev-72B-GPTQ-Int4: Quantization
- mlx-community/Kimi-Dev-72B-8bit: Quantization
- mlx-community/Kimi-Dev-72B-4bit: Quantization
- mlx-community/Kimi-Dev-72B-5bit: Quantization
- cnfusion/Kimi-Dev-72B-mlx-4Bit: Quantization
- mlx-community/Kimi-Dev-72B-6bit: Quantization
- unsloth/Kimi-Dev-72B: Identical copy
- cyankiwi/Kimi-Dev-72B-AWQ-4bit: Quantization
- ubergarm/Kimi-Dev-72B-GGUF: Quantization