yocabon commited on
Commit
42e0098
·
1 Parent(s): 35e2575

update instructions

Browse files
Files changed (1) hide show
  1. README.md +15 -4
README.md CHANGED
@@ -100,9 +100,8 @@ cd ../../../../
100
 
101
  ### Checkpoints
102
 
103
- TODO upload retrieval_model somewhere
104
-
105
- You can obtain the checkpoints by two ways:
106
 
107
  1) You can use our huggingface_hub integration: the models will be downloaded automatically.
108
 
@@ -124,6 +123,18 @@ wget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge
124
  For these checkpoints, make sure to agree to the license of all the training datasets we used, in addition to CC-BY-NC-SA 4.0.
125
  The mapfree dataset license in particular is very restrictive. For more information, check [CHECKPOINTS_NOTICE](CHECKPOINTS_NOTICE).
126
 
 
 
 
 
 
 
 
 
 
 
 
 
127
 
128
  ### Interactive demo
129
 
@@ -136,7 +147,7 @@ demo.py is the updated demo for MASt3R. It uses our new sparse global alignment
136
  python3 demo.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric
137
 
138
  # Use --weights to load a checkpoint from a local file, eg --weights checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth
139
- # Use --retrieval_model and point to the retrieval checkpoint to enable retrieval as a pairing strategy, asmk must be installed
140
  # Use --local_network to make it accessible on the local network, or --server_name to specify the url manually
141
  # Use --server_port to change the port, by default it will search for an available port starting at 7860
142
  # Use --device to use a different device, by default it's "cuda"
 
100
 
101
  ### Checkpoints
102
 
103
+ #### MASt3R Model
104
+ You can obtain the model checkpoints by two ways:
 
105
 
106
  1) You can use our huggingface_hub integration: the models will be downloaded automatically.
107
 
 
123
  For these checkpoints, make sure to agree to the license of all the training datasets we used, in addition to CC-BY-NC-SA 4.0.
124
  The mapfree dataset license in particular is very restrictive. For more information, check [CHECKPOINTS_NOTICE](CHECKPOINTS_NOTICE).
125
 
126
+ #### Retrieval Model
127
+ You need to download both the `trainingfree.pth` and `codebook.pkl` files, and put them in the same directory.
128
+
129
+ For `MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric`:
130
+ [`MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_trainingfree`](https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_trainingfree.pth)
131
+ [`MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_codebook`](https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_codebook.pkl)
132
+
133
+ ```bash
134
+ mkdir -p checkpoints/
135
+ wget https://download.europe.naverlabs.com/ComputerVision/MASt3R/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_trainingfree.pth -P checkpoints/
136
+ wget https://download.europe.naverlabs.com/ComputerVision/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric_retrieval_codebook.pkl -P checkpoints/
137
+ ```
138
 
139
  ### Interactive demo
140
 
 
147
  python3 demo.py --model_name MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric
148
 
149
  # Use --weights to load a checkpoint from a local file, eg --weights checkpoints/MASt3R_ViTLarge_BaseDecoder_512_catmlpdpt_metric.pth
150
+ # Use --retrieval_model and point to the retrieval checkpoint (*trainingfree.pth) to enable retrieval as a pairing strategy, asmk must be installed
151
  # Use --local_network to make it accessible on the local network, or --server_name to specify the url manually
152
  # Use --server_port to change the port, by default it will search for an available port starting at 7860
153
  # Use --device to use a different device, by default it's "cuda"