KeyArgo
/
smolagents
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #1 from ScientistIzaak/transformer-model-device 

Add device parameter for TransformerModel in models.py
This commit is contained in:
Izaak Curry 2025-01-02 21:56:22 -08:00 committed by GitHub
parent 6f87aee50e e2ac275d6e
commit 7dbddb9d7e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/smolagents/models.py
View File

@ -280,7 +280,14 @@ class HfApiModel(Model):
class TransformersModel(Model):
"""This engine initializes a model and tokenizer from the given `model_id`."""
"""This engine initializes a model and tokenizer from the given `model_id`.
Parameters:
model_id (`str`, *optional*, defaults to `"HuggingFaceTB/SmolLM2-1.7B-Instruct"`):
The Hugging Face model ID to be used for inference. This can be a path or model identifier from the Hugging Face model hub.
device (`str`, optional, defaults to `"cuda"` if available, else `"cpu"`.):
The device to load the model on (`"cpu"` or `"cuda"`).
"""
def __init__(self, model_id: Optional[str] = None, device: Optional[str] = None):
super().__init__()
@ -294,6 +301,7 @@ class TransformersModel(Model):
if device is None:
device = "cuda" if torch.cuda.is_available() else "cpu"
self.device = device
logger.info(f"Using device: {self.device}")
try:
self.tokenizer = AutoTokenizer.from_pretrained(model_id)
self.model = AutoModelForCausalLM.from_pretrained(model_id).to(self.device)
@ -302,7 +310,7 @@ class TransformersModel(Model):
f"Failed to load tokenizer and model for {model_id=}: {e}. Loading default tokenizer and model instead from {model_id=}."
)
self.tokenizer = AutoTokenizer.from_pretrained(default_model_id)
self.model = AutoModelForCausalLM.from_pretrained(default_model_id, device_map=device)
self.model = AutoModelForCausalLM.from_pretrained(default_model_id).to(self.device)
def make_stopping_criteria(self, stop_sequences: List[str]) -> StoppingCriteriaList:
class StopOnStrings(StoppingCriteria):