MathGenie/MathCoder2-CodeLlama-7B

Full fine-tune vs codellama/CodeLlama-7b-hf. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/1, Norms 65/65, MLP 96/96, Attention 128/128, lm_head 1/1

3 derivatives checked

• bartowski/MathCoder2-CodeLlama-7B-GGUF: Same weights, converted
• mradermacher/MathCoder2-CodeLlama-7B-i1-GGUF: Quantization
• mradermacher/MathCoder2-CodeLlama-7B-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse