Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed because of a cast error
Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 1 new columns ({'# EEG-EMG LOSO Pipeline (Deliverable)'}) and 1 missing columns ({'# Documentation: EEG-EMG LOSO Pipeline'}).

This happened while the csv dataset builder was generating data using

zip://teamspace/studios/this_studio/deliverable/README.md::hf://datasets/rverma0631/deliverable-dataset@9d33f2c2b451c72ad961323d3508c245d50afd30/deliverable.zip, ['hf://datasets/rverma0631/deliverable-dataset@9d33f2c2b451c72ad961323d3508c245d50afd30/deliverable.zip']

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1800, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 765, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2321, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2249, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              # EEG-EMG LOSO Pipeline (Deliverable): string
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 435
              to
              {'# Documentation: EEG-EMG LOSO Pipeline': Value('string')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1343, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 907, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1646, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1802, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 1 new columns ({'# EEG-EMG LOSO Pipeline (Deliverable)'}) and 1 missing columns ({'# Documentation: EEG-EMG LOSO Pipeline'}).
              
              This happened while the csv dataset builder was generating data using
              
              zip://teamspace/studios/this_studio/deliverable/README.md::hf://datasets/rverma0631/deliverable-dataset@9d33f2c2b451c72ad961323d3508c245d50afd30/deliverable.zip, ['hf://datasets/rverma0631/deliverable-dataset@9d33f2c2b451c72ad961323d3508c245d50afd30/deliverable.zip']
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

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.

# Documentation: EEG-EMG LOSO Pipeline
string
This document summarizes the pipeline architecture and reported results from the provided code in this deliverable.
## Pipeline overview
The pipeline runs in three stages:
1) Preprocess raw EEG/EMG HDF5 data and build a LOSO split.
2) Extract fixed-length epochs around go-cue events.
3) Train and evaluate models (v4 Mamba model and baselines).
## Architecture diagram
```mermaid
graph LR
A[Raw EEG HDF5] --> B[Preprocess: robust norm + band envelopes]
A2[Raw EMG HDF5] --> B2[Preprocess: robust norm]
B --> C[Train/Test split (LOSO)]
B2 --> C
C --> D[Epoch extraction around go-cue]
D --> E[EEG raw branch]
D --> F[EEG band branch]
E --> G[EEG embedding]
F --> G
G --> H[Backbone: BiMamba blocks + LocalMHA]
H --> I[Head]
I --> J[EMG regression output]
```
## Model details (v4)
- Inputs: raw EEG + EEG band envelopes
- Embedding: dual-branch (raw EEG projection + band EEG conv)
- Backbone: stacked BiMamba blocks with LocalMHA every N blocks
- Head: depthwise temporal conv + linear projection for EMG regression
## Results (from compare.txt)
LOSO subject: sub-8, EMG regression
Baseline comparison:
| Architecture | Params | MSE | Corr (all) | Corr (pre) | Corr (post) |
|---|---:|---:|---:|---:|---:|
| LSTM | 936,236 | 0.7490 | 0.2288 | -0.0121 | 0.1670 |
| BiLSTM | 862,508 | 0.6607 | 0.0114 | 0.0520 | 0.0008 |
| GRU | 861,740 | 0.6112 | 0.3746 | 0.0196 | 0.2916 |
| Transformer | 937,388 | 0.4324 | 0.5857 | 0.0260 | 0.5659 |
Our architecture (v4):
| Metric | Value |
|---|---:|
| MSE | 0.3857 |
| Corr (all) | 0.6267 |
| Corr (pre) | 0.0320 |
| Corr (post) | 0.6071 |
Per-channel correlation (v4):
| Channel | Corr |
|---|---:|
| DeltPost | 0.7338 |
| DeltAnt | 0.8234 |
| PectMaj | 0.8072 |
| TrapUp | 0.7829 |
| TrapMid | 0.0886 |
| TrapLow | 0.5513 |
| BicBrach | 0.7405 |
| TricLatH | 0.7076 |
| TricLong | 0.3609 |
| TricMedH | 0.5382 |
| Infraspin | 0.7595 |
## Reproducibility
- Results above are from the included run log in new_codes/compare.txt.
- Use the pinned uv.lock and Python 3.11/3.12 for consistent installs.
- See README.md for full setup and execution steps.
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
null
End of preview.

No dataset card yet

Downloads last month
10