Allow loading weights from local

chunked_pooling/wrappers.py

@@ -1,3 +1,4 @@
+import os
 from typing import List, Optional, Union
 
 import torch
@@ -138,7 +139,7 @@ def wrapper(self, *args, **kwargs):
     return wrapper
 
 
-def load_model(model_name, **model_kwargs):
+def load_model(model_name, model_weights=None, **model_kwargs):
     if model_name in MODEL_WRAPPERS:
         model = MODEL_WRAPPERS[model_name](model_name, **model_kwargs)
         if hasattr(MODEL_WRAPPERS[model_name], 'has_instructions'):
@@ -149,6 +150,9 @@ def load_model(model_name, **model_kwargs):
         model = AutoModel.from_pretrained(model_name, trust_remote_code=True)
         has_instructions = False
 
+    if model_weights and os.path.exists(model_weights):
+        model._model.load_state_dict(torch.load(model_weights, device=model.device))
+
     # encode functions of various models do not support all sentence transformers kwargs parameter
     if model_name in MODELS_WITHOUT_PROMPT_NAME_ARG:
         ENCODE_FUNC_NAMES = ['encode', 'encode_queries', 'encode_corpus']

run_chunked_eval.py

@@ -21,6 +21,11 @@
     default='jinaai/jina-embeddings-v2-small-en',
     help='The name of the model to use.',
 )
+@click.option(
+    '--model-weights',
+    default=None,
+    help='The path to the model weights to use, e.g. in case of finetuning.',
+)
 @click.option(
     '--strategy',
     default=DEFAULT_CHUNKING_STRATEGY,
@@ -70,6 +75,7 @@
 )
 def main(
     model_name,
+    model_weights,
     strategy,
     task_name,
     eval_split,
@@ -91,7 +97,7 @@ def main(
             f'Truncation is disabled because Long Late Chunking algorithm is enabled.'
         )
 
-    model, has_instructions = load_model(model_name)
+    model, has_instructions = load_model(model_name, model_weights)
 
     tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)