ARCHIVED TRAINING / facebookresearch/hydra
facebookresearch__hydra-273
This is a preserved training example. Membership in the current training wave has not been verified.
Problem statement
[Bug] handle yaml file with dot in the name
# 🐛 Bug
<!-- A clear and concise description of what the bug is. -->
When creating yaml file that has dot in the base name, `config_loader.py` erroneously misses the `.yaml` extension which results in a `hydra.errors.MissingConfigException`.
## To reproduce
configuration groups:
```
├── conf
│ └── lr
│ ├── 0.1.yaml
│ └── 0.01.yaml
└── my_app.py
```
with minimum `my_app.py` setup same as in the tutorial
```
python my_app.py lr=0.1
```
Traceback:
```
Traceback (most recent call last):
File "app.py", line 9, in <module>
app()
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/main.py", line 24, in decorated_main
strict=strict,
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/utils.py", line 56, in run_hydra
hydra.run(overrides=args.overrides)
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/hydra.py", line 103, in run
cfg = self._load_config(overrides)
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/hydra.py", line 345, in _load_config
cfg = self.config_loader.load_configuration(overrides)
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py", line 56, in load_configuration
cfg = self._merge_defaults(hydra_cfg, defaults, split_at)
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py", line 395, in _merge_defaults
cfg = merge_defaults(cfg, user_list)
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py", line 374, in merge_defaults
required=not is_optional,
File "/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py", line 301, in _merge_config
raise MissingConfigException(msg, new_cfg, options)
hydra.errors.MissingConfigException: Could not load lr/0.1, available options:
lr:
0.01
0.1
```
## Questions
- Currently, I have to edit `config_loader.py` file to add the `.yaml` extension whenever it is missing. Would this cause any unwanted behavior with other features?
- Is it generally not advisable to name config files with, say, float values? Are there better ways to incorporate float values in the config?
Training membership
opd_e4b_cispo_v5 opd_e4b_pure_v6a opd_e4b_pure_v6b opd_e4b_pure_v7 opd_e4b_pure_v7b opd_e4b_self_audit_v4
Task issues extracted from preserved verified_rft_messages training examples. This is not a verified current RL wave manifest.
Task and snapshot provenance
Snapshot captured 2026-11-12T17:06:59.476467+00:00 · Autoresearch-Arena/gemma-4-developer-agent
{
"datasets": [
"opd_e4b_cispo_v5",
"opd_e4b_pure_v6a",
"opd_e4b_pure_v6b",
"opd_e4b_pure_v7",
"opd_e4b_pure_v7b",
"opd_e4b_self_audit_v4"
],
"examples": 20,
"instance_id": "facebookresearch__hydra-273",
"problem_statement": "[Bug] handle yaml file with dot in the name\n# 🐛 Bug\r\n\r\n<!-- A clear and concise description of what the bug is. -->\r\nWhen creating yaml file that has dot in the base name, `config_loader.py` erroneously misses the `.yaml` extension which results in a `hydra.errors.MissingConfigException`.\r\n\r\n## To reproduce\r\nconfiguration groups:\r\n```\r\n├── conf\r\n│ └── lr\r\n│ ├── 0.1.yaml\r\n│ └── 0.01.yaml\r\n└── my_app.py\r\n```\r\nwith minimum `my_app.py` setup same as in the tutorial\r\n```\r\npython my_app.py lr=0.1\r\n```\r\nTraceback:\r\n```\r\nTraceback (most recent call last):\r\n File \"app.py\", line 9, in <module>\r\n app()\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/main.py\", line 24, in decorated_main\r\n strict=strict,\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/utils.py\", line 56, in run_hydra\r\n hydra.run(overrides=args.overrides)\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/hydra.py\", line 103, in run\r\n cfg = self._load_config(overrides)\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/hydra.py\", line 345, in _load_config\r\n cfg = self.config_loader.load_configuration(overrides)\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py\", line 56, in load_configuration\r\n cfg = self._merge_defaults(hydra_cfg, defaults, split_at)\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py\", line 395, in _merge_defaults\r\n cfg = merge_defaults(cfg, user_list)\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py\", line 374, in merge_defaults\r\n required=not is_optional,\r\n File \"/home/yjxiao/miniconda3/lib/python3.7/site-packages/hydra/_internal/config_loader.py\", line 301, in _merge_config\r\n raise MissingConfigException(msg, new_cfg, options)\r\nhydra.errors.MissingConfigException: Could not load lr/0.1, available options:\r\nlr:\r\n\t0.01\r\n\t0.1\r\n```\r\n## Questions\r\n- Currently, I have to edit `config_loader.py` file to add the `.yaml` extension whenever it is missing. Would this cause any unwanted behavior with other features? \r\n- Is it generally not advisable to name config files with, say, float values? Are there better ways to incorporate float values in the config?",
"repo": "facebookresearch/hydra",
"source": "v2/experiments/train_data/opd_e4b_cispo_v5/verified_rft_messages.jsonl",
"source_sha256": "a43610ad0dfb6ace95375cfae09000f41afc0338cc925af870a4ade7a9b20c4c"
}{
"captured_at": "2026-11-12T17:06:59.476467+00:00",
"membership": "archived training",
"membership_evidence": "Task issues extracted from preserved verified_rft_messages training examples. This is not a verified current RL wave manifest.",
"name": "Preserved post-training task archive",
"source": "Autoresearch-Arena/gemma-4-developer-agent",
"source_revision": "016496de222aaafca0d4ee4c18efc63ba616f76e",
"sources": [
{
"path": "v2/experiments/train_data/opd_e4b_cispo_v5/verified_rft_messages.jsonl",
"sha256": "a43610ad0dfb6ace95375cfae09000f41afc0338cc925af870a4ade7a9b20c4c"
},
{
"path": "v2/experiments/train_data/opd_e4b_hybrid_v3/verified_rft_messages.jsonl",
"sha256": "bd326303eb127e8e5e08f03daf97bb88d79d19709d59f3f93f348bc56686a8df"
},
{
"path": "v2/experiments/train_data/opd_e4b_onpolicy_v3/verified_rft_messages.jsonl",
"sha256": "960c680e797e823a0d3a22f4b1d5ef60a72971373d703b877340b0c9126aaf68"
},
{
"path": "v2/experiments/train_data/opd_e4b_pure_v6a/verified_rft_messages.jsonl",
"sha256": "f045b7ca5b3883a1e2eaee81e1c02a4a1b53dd7b7b36e98bacb2ba5823a15148"
},
{
"path": "v2/experiments/train_data/opd_e4b_pure_v6b/verified_rft_messages.jsonl",
"sha256": "0f62460a9706eea5e1605255896f191ab00cb5706c5f1f2e1c293e9a718218dc"
},
{
"path": "v2/experiments/train_data/opd_e4b_pure_v7/verified_rft_messages.jsonl",
"sha256": "e37f7bd675322f3f81500b73559d60d5f602ca0537bf19f5cb06332c728c7b72"
},
{
"path": "v2/experiments/train_data/opd_e4b_pure_v7b/verified_rft_messages.jsonl",
"sha256": "b864bdc8695183433eef6e31cb8d33fad44ba3aded2170c1ade4ad5b2944ce23"
},
{
"path": "v2/experiments/train_data/opd_e4b_self_audit_v4/verified_rft_messages.jsonl",
"sha256": "0026023745afb8f3dc1c176dc7f7171157b4ad5a66656b524af218e8e19ceee6"
},
{
"path": "v2/experiments/train_data/rft_e4b_aligned_v2/verified_rft_messages.jsonl",
"sha256": "92a4515fee9d152b0d4c510ee3ce4caaaffe2003ba7584469f8027f174aab18c"
},
{
"path": "v2/experiments/train_data/rft_v1/verified_rft_messages.jsonl",
"sha256": "a42453ccd7ac7680b0a03865e8e7080bc8bf74c8d9d193b440c441ed46f8512b"
},
{
"path": "v2/experiments/train_data/rft_v2/verified_rft_messages.jsonl",
"sha256": "18f4b3bd7e651f83c6acab479c9475e3fa50dc652a384f309e97146983ac85ba"
}
]
}