DavidAU/Qwen3-Coder-53B-A3B-Instruct-TOTAL-RECALL-v2-MASTER-CODER-L

Partial fine-tune (fewer layers) vs Qwen/Qwen3-Coder-30B-A3B-Instruct. Qwen/Qwen3-Coder-30B-A3B-Instruct matches part of DavidAU/Qwen3-Coder-53B-A3B-Instruct-TOTAL-RECALL-v2-MASTER-CODER-L (18867 of its 33015 tensors; the rest of DavidAU/Qwen3-Coder-53B-A3B-Instruct-TOTAL-RECALL-v2-MASTER-CODER-L has no counterpart). On the shared part: Partial change: 11% of the weights changed (attention, lm_head, mlp, norms), in 1 of 48 layers.

Tensors changed: Embeddings 0/1, Norms 5/193, MLP 30/1337, Attention 4/192, lm_head 1/1

2 derivatives checked

• mradermacher/Qwen3-Coder-53B-A3B-Instruct-TOTAL-RECALL-v2-MASTER-CODER-L-i1-GGUF: Quantization
• mradermacher/Qwen3-Coder-53B-A3B-Instruct-TOTAL-RECALL-v2-MASTER-CODER-L-GGUF: Quantization

On the Hub · Downloads on Model Pulse