CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-30 08:42:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

convert.py : fix HF tensor permuting / unpacking

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2023-08-17 21:06:45 +03:00
parent 78e1e57862
commit acaa98234a
2 changed files with 24 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
convert.py
View File

@@ -812,6 +812,23 @@ def convert_to_output_type(model: LazyModel, output_type: GGMLFileType) -> LazyM
def convert_model_names(model: LazyModel, params: Params) -> LazyModel:
tmap = gguf.get_tensor_name_map(ARCH, params.n_layer)
tmp = model
# HF models permut or pack some of the tensors, so we need to undo that
for i in itertools.count():
if f"model.layers.{i}.self_attn.q_proj.weight" in model:
print(f"Permuting layer {i}")
tmp[f"model.layers.{i}.self_attn.q_proj.weight"] = permute_lazy(model[f"model.layers.{i}.self_attn.q_proj.weight"], params.n_head, params.n_head_kv)
tmp[f"model.layers.{i}.self_attn.k_proj.weight"] = permute_lazy(model[f"model.layers.{i}.self_attn.k_proj.weight"], params.n_head, params.n_head_kv)
#tmp[f"model.layers.{i}.self_attn.v_proj.weight"] = model[f"model.layers.{i}.self_attn.v_proj.weight"]
elif f"model.layers.{i}.self_attn.W_pack.weight" in model:
print(f"Unpacking and permuting layer {i}")
tmp[f"model.layers.{i}.self_attn.q_proj.weight"] = permute_part_lazy(model[f"model.layers.{i}.self_attn.W_pack.weight"], 0, params.n_head, params.n_head_kv)
tmp[f"model.layers.{i}.self_attn.k_proj.weight"] = permute_part_lazy(model[f"model.layers.{i}.self_attn.W_pack.weight"], 1, params.n_head, params.n_head_kv)
tmp[f"model.layers.{i}.self_attn.v_proj.weight"] = part_lazy (model[f"model.layers.{i}.self_attn.W_pack.weight"], 2)
else:
break
out: LazyModel = {}
for name, lazy_tensor in model.items():
name_new = name
@@ -825,8 +842,9 @@ def convert_model_names(model: LazyModel, params: Params) -> LazyModel:
else:
raise Exception(f"Unexpected tensor name: {name}")
if gguf.should_skip_tensor(ARCH, params.n_layer, name_new):
if gguf.should_skip_tensor_TMP(ARCH, params.n_layer, name_new):
print(f"skipping tensor {name_new}")
continue
else:
print(f"{name:48s} -> {name_new:40s} | {lazy_tensor.data_type} | {lazy_tensor.shape}")
out[name_new] = lazy_tensor