vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fixes for model conversion

Browse source 
Co-authored-by: FNsi <125447286+FNsi@users.noreply.github.com>
This commit is contained in:
Henri Vasserman 2023-05-25 12:15:33 +03:00
parent ff99507049
commit aeacc57d3b
No known key found for this signature in database
GPG key ID: 2995FC0F58B1A986
1 changed files with 4 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

6
convert.py
View file

@ -149,7 +149,7 @@ class Params:
# TODO: hack for open_llama_3b
if n_embd == 3200:
n_mult = 108
n_mult = 216
n_head = 32
n_layer = 26
@ -607,7 +607,9 @@ def convert_transformers_to_orig(model: LazyModel) -> LazyModel:
out["norm.weight"] = model["model.norm.weight"]
out["output.weight"] = model["lm_head.weight"]
n_head = model["model.layers.0.self_attn.q_proj.weight"].shape[1] // 128
# TODO: hack for open_llama_3b
n_embd = model["model.layers.0.self_attn.q_proj.weight"].shape[1]
n_head = 32 if n_embd == 3200 else n_embd // 128
for i in itertools.count():
if f"model.layers.{i}.self_attn.q_proj.weight" not in model:
break