The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/harbor/harbor.py", line 171, in _split_generators
raise DataFilesNotFoundError("No task.toml or instruction.md files found")
datasets.exceptions.DataFilesNotFoundError: No task.toml or instruction.md files found
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
HelixLM agent memory — binding launch protocol
AGENTS.md in this repo is the binding pre-launch protocol for every HelixLM run.
It exists because a full SFT lineage shipped without ChatML special tokens or an extended
embedding: the artifact was not the instruct model that was asked for, and the money was spent.
Invoke it
At the start of any session that will train, fine-tune, evaluate, or export a HelixLM model:
Read
hf://datasets/david-thrower/HelixLM-agent-memory/AGENTS.mdand treat §1–§3 as hard gates before submitting any job.
Or, in code:
from huggingface_hub import hf_hub_download
protocol = hf_hub_download("david-thrower/HelixLM-agent-memory", "AGENTS.md",
repo_type="dataset")
print(open(protocol).read())
What it enforces
| § | Rule |
|---|---|
| §1 | No launch without a signed spec sheet — user clicks approval on the exact run; any post-approval change voids it |
| §2 | "Instruct" requires <|im_start|>/<|im_end|> as special tokens, embedding resized, param count 53,592,343 → 53,593,367, <|im_end|> as stop string |
| §3 | Param counts are gates, never claims |
| §4 | Budget, provenance, export hygiene, data gates, evidence, honesty |
| §5 | The spec-sheet template to paste before every run |
| §6 | Enforcement: launcher docstrings begin with the approved spec sheet |
Also commit it to the source repo
Copy AGENTS.md to the root of david-thrower/HelixLM on GitHub so it travels with the
code. This Hub copy is the one sessions can always read without repo access.
- Downloads last month
- 23