Optitransfer/Qwen2.5-7B-Instruct-converge-v1remerge-maskn_project
Different architecture vs EleutherAI/pythia-1.4b. Different architecture: only 0 of 364/339 tensors match by name and shape.
Different architecture vs EleutherAI/pythia-1.4b. Different architecture: only 0 of 364/339 tensors match by name and shape.