Skip to content

Commit

Permalink
Cleanup of prepare-training-data-bert
Browse files Browse the repository at this point in the history
  • Loading branch information
arjunsuresh committed Jun 12, 2023
1 parent 30fe073 commit 26b7a5f
Show file tree
Hide file tree
Showing 4 changed files with 22 additions and 7 deletions.
1 change: 0 additions & 1 deletion cm-mlops/automation/script/module.py
Original file line number Diff line number Diff line change
Expand Up @@ -2360,7 +2360,6 @@ def _run_deps(self, deps, clean_env_keys_deps, env, state, const, const_state, a
d['tags']+=","+t+env[key]

run_state['deps'].append(d['tags'])
run_state['deps'].append(d['tags'])

if not run_state['fake_deps']:
import copy
Expand Down
1 change: 1 addition & 0 deletions cm-mlops/script/download-file/_cm.json
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@
"automation_alias": "script",
"automation_uid": "5b4e0237da074764",
"cache": false,
"can_force_cache": true,
"deps": [],
"env": {
},
Expand Down
1 change: 1 addition & 0 deletions cm-mlops/script/extract-file/_cm.json
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@
}
],
"cache": false,
"can_force_cache": true,
"env": {
"CM_DAE_REMOVE_EXTRACTED": "yes"
},
Expand Down
26 changes: 20 additions & 6 deletions cm-mlops/script/prepare-training-data-bert/_cm.json
Original file line number Diff line number Diff line change
Expand Up @@ -36,25 +36,31 @@
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1fbGClQMi2CoMv7fwrwTC5YYPooQBdcFW",
"env": {
"CM_DOWNLOAD_FILENAME": "bert_config.json",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CONFIG_DOWNLOAD_DIR>>>"
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CONFIG_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "7f59165e21b7d566db610ff6756c926b"
},
"force_cache": true,
"extra_cache_tags": "mlperf,training,bert,config"
},
{
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1USK108J6hMM_d27xCHi738qBL8_BT1u1",
"env": {
"CM_DOWNLOAD_FILENAME": "vocab.txt",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_VOCAB_DOWNLOAD_DIR>>>"
"CM_DOWNLOAD_PATH": "<<<CM_BERT_VOCAB_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "64800d5d8528ce344256daf115d4965e"
},
"force_cache": true,
"extra_cache_tags": "bert,vocab"
},
{
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1tmMgLwoBvbEJEHXh77sqrXYw5RpqT8R_",
"env": {
"CM_DOWNLOAD_FILENAME": "bert_reference_results_text_md5.txt",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_DATA_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_FINAL_ENV_NAME": "CM_BERT_REFERENCE_RESULTS_TEXT_MD5_FILE_PATH"
"CM_DOWNLOAD_FINAL_ENV_NAME": "CM_BERT_REFERENCE_RESULTS_TEXT_MD5_FILE_PATH",
"CM_DOWNLOAD_CHECKSUM": "7d3a0619cb8bf7e829af99fa5c29daa8"
},
"force_cache": true,
"extra_cache_tags": "bert,data,results,md5"
},
{
Expand All @@ -64,32 +70,40 @@
"CM_EXTRACT_EXTRACTED_FILENAME": "results4",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_DATA_DOWNLOAD_DIR>>>",
"CM_EXTRACT_PATH": "<<<CM_BERT_DATA_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "",
"CM_EXTRACT_EXTRACTED_CHECKSUM_FILE": "<<<CM_BERT_REFERENCE_RESULTS_TEXT_MD5_FILE_PATH>>>"
},
"force_cache": true,
"extra_cache_tags": "bert,data,results"
},
{
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1chiTBljF0Eh1U5pKs6ureVHgSbtU8OG_",
"env": {
"CM_DOWNLOAD_FILENAME": "model.ckpt-28252.data-00000-of-00001",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>"
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "50797acd537880bfb5a7ade80d976129"
},
"force_cache": true,
"extra_cache_tags": "bert,checkpoint,data"
},
{
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1Q47V3K3jFRkbJ2zGCrKkKk-n0fvMZsa0",
"env": {
"CM_DOWNLOAD_FILENAME": "model.ckpt-28252.index",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>"
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "f97de3ae180eb8d479555c939d50d048"
},
"force_cache": true,
"extra_cache_tags": "bert,checkpoint,index"
},
{
"tags": "download,file,_gdown,_url.https://drive.google.com/uc?id=1vAcVmXSLsLeQ1q7gvHnQUSth5W_f_pwv",
"env": {
"CM_DOWNLOAD_FILENAME": "model.ckpt-28252.meta",
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>"
"CM_DOWNLOAD_PATH": "<<<CM_BERT_CHECKPOINT_DOWNLOAD_DIR>>>",
"CM_DOWNLOAD_CHECKSUM": "dbd16c731e8a8113bc08eeed0326b8e7"
},
"force_cache": true,
"extra_cache_tags": "bert,checkpoint,meta"
}
],
Expand Down

0 comments on commit 26b7a5f

Please sign in to comment.