Hugging Face's logo

vidore
/
colSmol-256M

ColPali
Safetensors
English
idefics3
colsmolvlm
vidore-experimental
vidore
Model card Files Community
manu commited on
Commit
c2790ac
·
verified ·
1 Parent(s): 88b54c6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -3
README.md CHANGED
@@ -1,7 +1,7 @@
1
  ---
2
  license: mit
3
  library_name: colpali
4
- base_model: vidore/ColSmolVLM-Instruct-250M
5
  language:
6
  - en
7
  tags:
@@ -9,7 +9,7 @@ tags:
9
  - vidore-experimental
10
  - vidore
11
  ---
12
- # ColSmolVLM-Instruct-250M: Visual Retriever based on SmolVLM-Instruct-250M with ColBERT strategy
13
 
14
  ### This is a version trained with batch_size 32 for 3 epochs
15
 
@@ -58,7 +58,7 @@ from PIL import Image
58
  from colpali_engine.models import ColIdefics3, ColIdefics3Processor
59
 
60
  model = ColIdefics3.from_pretrained(
61
- "vidore/colSmol-250M",
62
  torch_dtype=torch.bfloat16,
63
  device_map="cuda:0",
64
  attn_implementation="flash_attention_2" # or eager
 
1
  ---
2
  license: mit
3
  library_name: colpali
4
+ base_model: vidore/ColSmolVLM-Instruct-256M
5
  language:
6
  - en
7
  tags:
 
9
  - vidore-experimental
10
  - vidore
11
  ---
12
+ # ColSmolVLM-Instruct-256M: Visual Retriever based on SmolVLM-Instruct-250M with ColBERT strategy
13
 
14
  ### This is a version trained with batch_size 32 for 3 epochs
15
 
 
58
  from colpali_engine.models import ColIdefics3, ColIdefics3Processor
59
 
60
  model = ColIdefics3.from_pretrained(
61
+ "vidore/colSmol-256M",
62
  torch_dtype=torch.bfloat16,
63
  device_map="cuda:0",
64
  attn_implementation="flash_attention_2" # or eager