zed-industries/zeta-2
LoRA-style vs ByteDance-Seed/Seed-Coder-8B-Base. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/65, MLP 96/96, Attention 128/128, lm_head 0/1
5 derivatives checked
- bartowski/zed-industries_zeta-2-GGUF: Same weights, converted
- bluevoid-pl/zeta2-GGUF: Same weights, converted
- NexVeridian/zeta-2-4bit: Quantization
- liodon-ai/zeta-2-imatrix-GGUF: Quantization
- vilarin/zeta-2-GGUF: Same weights, converted