CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-03 09:22:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

convert-llama-h5-to-gguf.py : add tensor data layout

Browse Source
This commit is contained in:
klosax
2023-08-15 19:54:33 +02:00
committed by GitHub
parent b6056c3db8
commit 66756c82af
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
convert-llama-h5-to-gguf.py
View File

@@ -97,6 +97,7 @@ gguf_writer.add_architecture(llm_arch)
gguf_writer.add_name(last_dir) gguf_writer.add_name(last_dir)
gguf_writer.add_file_type("All tensors F32" if ftype == 0 else "Most tensors F16, some F32") gguf_writer.add_file_type("All tensors F32" if ftype == 0 else "Most tensors F16, some F32")
gguf_writer.add_source_hf_repo(hf_repo) gguf_writer.add_source_hf_repo(hf_repo)
gguf_writer.add_tensor_data_layout(llm_arch, "Meta AI original pth")
gguf_writer.add_context_length(llm_arch, hparams["max_position_embeddings"]) gguf_writer.add_context_length(llm_arch, hparams["max_position_embeddings"])
gguf_writer.add_embedding_length(llm_arch, hparams["hidden_size"]) gguf_writer.add_embedding_length(llm_arch, hparams["hidden_size"])
gguf_writer.add_block_count(llm_arch, block_count) gguf_writer.add_block_count(llm_arch, block_count)