nv-bschifferer commited on
Commit
1f0fdea
·
1 Parent(s): 6a21313

update readme

Browse files
Files changed (2) hide show
  1. README.md +3 -4
  2. config.json +0 -1
README.md CHANGED
@@ -20,13 +20,11 @@ pipeline_tag: visual-document-retrieval
20
 
21
  # llama-nemoretriever-colembed-1b-v1
22
 
23
- # llama-nemoretriever-colembed-1b-v1
24
-
25
  ## Description
26
 
27
  The **nvidia/llama-nemoretriever-colembed-1b-v1** is a late interaction embedding model fine-tuned for query-document retrieval. Users can input `queries`, which are text, or `documents` which are page images, to the model. The model outputs ColBERT-style multi-vector numerical representations for input queries and documents. It is the smaller version of [llama-nemoretriever-colembed-3b-v1](https://huggingface.co/nvidia/llama-nemoretriever-colembed-3b-v1), which achieved 1st place on ViDoRe V1 (nDCG@5), ViDoRe V2 (nDCG@5) and MTEB VisualDocumentRetrieval (Rank Borda) (as of 27th June, 2025). **nvidia/llama-nemoretriever-colembed-1b-v1** achieves 2nd place on the benchmarks.
28
 
29
- This model is for non-commercial/research use only. |
30
 
31
  ### License/Terms of Use
32
  Governing Terms: [NVIDIA License](https://huggingface.co/nvidia/llama-nemoretriever-colembed-1b-v1/blob/main/LICENSE)
@@ -114,11 +112,12 @@ from transformers import AutoModel
114
 
115
  # Load Model
116
  model = AutoModel.from_pretrained(
117
- 'nvidia/llama-NemoRetriever-ColEmbed-1B-v1',
118
  device_map='cuda',
119
  trust_remote_code=True,
120
  torch_dtype=torch.bfloat16,
121
  attn_implementation="flash_attention_2",
 
122
  ).eval()
123
 
124
  # Queries
 
20
 
21
  # llama-nemoretriever-colembed-1b-v1
22
 
 
 
23
  ## Description
24
 
25
  The **nvidia/llama-nemoretriever-colembed-1b-v1** is a late interaction embedding model fine-tuned for query-document retrieval. Users can input `queries`, which are text, or `documents` which are page images, to the model. The model outputs ColBERT-style multi-vector numerical representations for input queries and documents. It is the smaller version of [llama-nemoretriever-colembed-3b-v1](https://huggingface.co/nvidia/llama-nemoretriever-colembed-3b-v1), which achieved 1st place on ViDoRe V1 (nDCG@5), ViDoRe V2 (nDCG@5) and MTEB VisualDocumentRetrieval (Rank Borda) (as of 27th June, 2025). **nvidia/llama-nemoretriever-colembed-1b-v1** achieves 2nd place on the benchmarks.
26
 
27
+ This model is for non-commercial/research use only.
28
 
29
  ### License/Terms of Use
30
  Governing Terms: [NVIDIA License](https://huggingface.co/nvidia/llama-nemoretriever-colembed-1b-v1/blob/main/LICENSE)
 
112
 
113
  # Load Model
114
  model = AutoModel.from_pretrained(
115
+ 'nvidia/llama-nemoretriever-colembed-1b-v1',
116
  device_map='cuda',
117
  trust_remote_code=True,
118
  torch_dtype=torch.bfloat16,
119
  attn_implementation="flash_attention_2",
120
+ revision='6a21313a150a903bc522dc0d15ed47784a0d4c8d'
121
  ).eval()
122
 
123
  # Queries
config.json CHANGED
@@ -1,6 +1,5 @@
1
  {
2
  "_commit_hash": null,
3
- "_name_or_path": "./model_1b_test/",
4
  "architectures": [
5
  "llama_NemoRetrieverColEmbed"
6
  ],
 
1
  {
2
  "_commit_hash": null,
 
3
  "architectures": [
4
  "llama_NemoRetrieverColEmbed"
5
  ],