mostly replaced by decoding tensors except ggml models which only support llama
Attach a file by drag & drop or click to upload
