TinyMyo: Update hand gesture classification experiment details - #8

Open

MatteoFasulo wants to merge 24 commits into

pulp-bio:mainfrom

MatteoFasulo:patch-1

Commits on Jul 27, 2026

Update hand gesture classification experiment details
MatteoFasulo
committed
1802ea7
Clarify dataset and performance metrics in TinyMyo.md
MatteoFasulo
committed
d591309
feat: Add optional Weights & Biases logger integration to training script
MatteoFasulo
committed
fbb4dc9
feat: Add local dataset implementations module for BioFoundation. This avoids namespace issues with third party libraries (e.g., datasets from HuggingFace)
MatteoFasulo
committed
73bd396
fix: Comment out Grouped Query Attention (GQA) for backward compatibility with old pytorch versions (<2.5)
MatteoFasulo
committed
0d0aa83
feat: EMG dataset handling - large scale pretraining using shared memory optimization for DDP
MatteoFasulo
committed
1604197
feat: add TinyMyo finetuning and pretraining configurations for 4 layers, EMG finetuning dataset handling, and improved training script logging
MatteoFasulo
committed
3dbcaac
refactor: training script with wandb offline parameter and finish closure
MatteoFasulo
committed
7519a77
refactor: TinyMyo configuration and model for improved training efficiency and logging
MatteoFasulo
committed
1354e4e
Update TinyMyo with safe asserts for number of channels, per-head dim, and sequence length
MatteoFasulo
committed
b133d05
refactor: removed unused mask token during TinyMyo finetuning
MatteoFasulo
committed
439e5f0
refactor: removed black and isort configurations from project file
MatteoFasulo
committed
81e1545
Removed init.py from datasets folder (introduced to avoid namespace issues with huggingface datasets)
MatteoFasulo
committed
ece5874
training_step and validation_step now call as_signal_batch directly
MatteoFasulo
committed
035bc35
Update TinyMyo documentation
MatteoFasulo
committed
261adb1
docs: update TinyMyo documentation for clarity and structure
MatteoFasulo
committed
e9f95f5
docs: enhance TinyMyo documentation with classification checkpoints and details
MatteoFasulo
committed
927e357
Removed pyproject.toml . Stick to requirements.txt for repo deps
MatteoFasulo
committed
9aff364

Commits on Jul 28, 2026

Commits on Jul 29, 2026

Update TinyMyo fine-tuning configuration and documentation
MatteoFasulo
committed
a259920

Commits on Jul 31, 2026

Update TinyMyo pretraining configuration and dataset handling
MatteoFasulo
committed
842cade

Navigation Menu

Uh oh!

TinyMyo: Update hand gesture classification experiment details - #8

TinyMyo: Update hand gesture classification experiment details#8
MatteoFasulo wants to merge 24 commits into
pulp-bio:mainfrom
MatteoFasulo:patch-1

Commits on Jul 27, 2026

Update hand gesture classification experiment details

Clarify dataset and performance metrics in TinyMyo.md

feat: Add optional Weights & Biases logger integration to training script

feat: Add local dataset implementations module for BioFoundation. This avoids namespace issues with third party libraries (e.g., datasets from HuggingFace)

fix: Comment out Grouped Query Attention (GQA) for backward compatibility with old pytorch versions (<2.5)

feat: EMG dataset handling - large scale pretraining using shared memory optimization for DDP

feat: add TinyMyo finetuning and pretraining configurations for 4 layers, EMG finetuning dataset handling, and improved training script logging

refactor: training script with wandb offline parameter and finish closure

refactor: TinyMyo configuration and model for improved training efficiency and logging

Update TinyMyo with safe asserts for number of channels, per-head dim, and sequence length

refactor: removed unused mask token during TinyMyo finetuning

refactor: removed black and isort configurations from project file

Removed init.py from datasets folder (introduced to avoid namespace issues with huggingface datasets)

training_step and validation_step now call as_signal_batch directly

Update TinyMyo documentation

docs: update TinyMyo documentation for clarity and structure

docs: enhance TinyMyo documentation with classification checkpoints and details

Removed pyproject.toml . Stick to requirements.txt for repo deps

Commits on Jul 28, 2026

Update fine-tuning EMG task definition

Update fine-tuning EMG training metrics logging

Update fine-tuning EMG task with correct scheduler steps

Update EMG tasks with correct scheduler steps and LR-decay

Commits on Jul 29, 2026

Update TinyMyo fine-tuning configuration and documentation

Commits on Jul 31, 2026

Update TinyMyo pretraining configuration and dataset handling