diffusion-gif commited on
Commit
c64b064
·
verified ·
1 Parent(s): ac78ced

data: attach pinned training dataset index and safe recovery helpers

Browse files
Files changed (3) hide show
  1. CHECKSUMS.sha256 +2 -2
  2. README.md +1 -1
  3. bundle_manifest.json +3 -3
CHECKSUMS.sha256 CHANGED
@@ -1,4 +1,4 @@
1
- de97f95cc56c1400755975a453d1da522b15614bf92e298a23e237fe1d9aa14f README.md
2
  4e0e37529fbf86ebef6bf7cadec4dda1b762f31ed6dba7815a79732bc6b7e836 checkpoints/caption_grpo/adapter_config.json
3
  290aacd2808639e3c3e10c81c69acfdd843100b8888ad183a2e03ba6e43f5d11 checkpoints/caption_grpo/adapter_model.safetensors
4
  62671bc8d430001734a90c33c0059e563388eba42706a5c8603704018035ae8b checkpoints/caption_grpo/additional_config.json
@@ -80,4 +80,4 @@ c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4 LICENSES/Apach
80
  2569cdcdb30128ffb7d1f178a2669580ead36e532f9f8c65a67f0f25c543caaf TRAINING_DATA_INDEX.json
81
  67acb0f58a052ce7fd309e50cdaf473a9880bc31a8d1c7f68212d528088d275c tools/fetch_training_data.py
82
  4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826 tools/restore_legacy_mix6.py
83
- f495623211474f8cd4f76f07abeb24275e3b743bdeb7e09f5bfde3d54f85bbc3 bundle_manifest.json
 
1
+ 6b1b91e8bd12905d33207cea6ea84e8c651517f26abf790f899f7fb6160aebf4 README.md
2
  4e0e37529fbf86ebef6bf7cadec4dda1b762f31ed6dba7815a79732bc6b7e836 checkpoints/caption_grpo/adapter_config.json
3
  290aacd2808639e3c3e10c81c69acfdd843100b8888ad183a2e03ba6e43f5d11 checkpoints/caption_grpo/adapter_model.safetensors
4
  62671bc8d430001734a90c33c0059e563388eba42706a5c8603704018035ae8b checkpoints/caption_grpo/additional_config.json
 
80
  2569cdcdb30128ffb7d1f178a2669580ead36e532f9f8c65a67f0f25c543caaf TRAINING_DATA_INDEX.json
81
  67acb0f58a052ce7fd309e50cdaf473a9880bc31a8d1c7f68212d528088d275c tools/fetch_training_data.py
82
  4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826 tools/restore_legacy_mix6.py
83
+ 7ca67b923529b5bffcb1d9cef166e1a12439f0156720b3d8044966c2dd022115 bundle_manifest.json
README.md CHANGED
@@ -118,7 +118,7 @@ Public registry: https://huggingface.co/datasets/diffusion-gif/training-data-ind
118
 
119
  ```bash
120
  # For old trajectory checkpoints; inspect other dataset IDs for Any2Ego:
121
- python tools/fetch_training_data.py --index TRAINING_DATA_INDEX.json --dataset legacy_mix6_sft --output data_download
122
  # after owner login / dry-run inspection add --execute
123
  ```
124
 
 
118
 
119
  ```bash
120
  # For old trajectory checkpoints; inspect other dataset IDs for Any2Ego:
121
+ python tools/fetch_training_data.py --index TRAINING_DATA_INDEX.json --dataset ego_exo_original_training_inputs --output data_download
122
  # after owner login / dry-run inspection add --execute
123
  ```
124
 
bundle_manifest.json CHANGED
@@ -176,8 +176,8 @@
176
  "files": [
177
  {
178
  "path": "README.md",
179
- "bytes": 7610,
180
- "sha256": "de97f95cc56c1400755975a453d1da522b15614bf92e298a23e237fe1d9aa14f"
181
  },
182
  {
183
  "path": "checkpoints/caption_grpo/adapter_config.json",
@@ -585,7 +585,7 @@
585
  "sha256": "4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826"
586
  }
587
  ],
588
- "total_file_bytes": 46092647806,
589
  "largest_file_bytes": 4999831048,
590
  "native_reader": "safetensors.safe_open(framework=pt, device=cpu), slices only",
591
  "validation": {
 
176
  "files": [
177
  {
178
  "path": "README.md",
179
+ "bytes": 7627,
180
+ "sha256": "6b1b91e8bd12905d33207cea6ea84e8c651517f26abf790f899f7fb6160aebf4"
181
  },
182
  {
183
  "path": "checkpoints/caption_grpo/adapter_config.json",
 
585
  "sha256": "4cd8ba59a93ab631419f29b61153ab8024cc7392f81fca1e0470c400144c8826"
586
  }
587
  ],
588
+ "total_file_bytes": 46092647823,
589
  "largest_file_bytes": 4999831048,
590
  "native_reader": "safetensors.safe_open(framework=pt, device=cpu), slices only",
591
  "validation": {