tacos4me/GLM-5.3-Flash-NVFP4-FP8ATTN-512K
Partial fine-tune vs LibertAIDAI/GLM-5.3-Flash-NVFP4. Partial change: 84% of the weights changed (attention, lm_head, mlp, other), in 46 of 46 layers.
Tensors changed: Attention 334/827, MLP 213/1339, lm_head 1/1, Embeddings 0/3, Norms 0/277, other 1/4