Skip to content

Fix ESMC 300M/600M local loading: torch.load the .pth (repos ship no safetensors) #332

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
sakzgupzzz wants to merge 1 commit into
base: main
Choose a base branch
from
+16 −2
Open

Fix ESMC 300M/600M local loading: torch.load the .pth (repos ship no safetensors) #332

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 16 additions & 2 deletions esm/pretrained.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,7 +72,14 @@ def ESMC_300M_202412(device: torch.device | str = "cpu", use_flash_attn: bool =
tokenizer=get_esmc_model_tokenizers(),
use_flash_attn=use_flash_attn,
).eval()
load_torch_model(model, data_root("esmc-300"))
# The 300M repo ships a single legacy .pth (no safetensors), so load it
# explicitly like the ESM3 loaders above rather than via load_torch_model,
# which (safe=True) expects a safetensors checkpoint in the directory.
state_dict = torch.load(
data_root("esmc-300") / "data/weights/esmc_300m_2024_12_v0.pth",
map_location=device,
)
model.load_state_dict(state_dict, assign=True)
model = model.to(device)
return model

Expand All @@ -86,7 +93,14 @@ def ESMC_600M_202412(device: torch.device | str = "cpu", use_flash_attn: bool =
tokenizer=get_esmc_model_tokenizers(),
use_flash_attn=use_flash_attn,
).eval()
load_torch_model(model, data_root("esmc-600"))
# The 600M repo ships a single legacy .pth (no safetensors), so load it
# explicitly like the ESM3 loaders above rather than via load_torch_model,
# which (safe=True) expects a safetensors checkpoint in the directory.
state_dict = torch.load(
data_root("esmc-600") / "data/weights/esmc_600m_2024_12_v0.pth",
map_location=device,
)
model.load_state_dict(state_dict, assign=True)
model = model.to(device)
return model

Expand Down