Dataset Viewer
Auto-converted to Parquet Duplicate
train_micro_batch_size_per_gpu
int64
1
8
gradient_accumulation_steps
int64
1
1
gradient_clipping
float64
1
1
steps_per_print
int64
1
25
zero_optimization
dict
bf16
dict
optimizer
dict
activation_checkpointing
dict
wall_clock_breakdown
bool
1 class
1
1
1
1
{ "stage": 2, "offload_optimizer": { "device": "none" }, "contiguous_gradients": true, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 50000000, "allgather_bucket_size": 50000000 }
{ "enabled": true }
{ "type": "AdamW", "params": { "lr": 0.00005, "betas": [ 0.9, 0.999 ], "eps": 0.000001, "weight_decay": 0 } }
{ "partition_activations": false, "cpu_checkpointing": false, "contiguous_memory_optimization": false, "number_checkpoints": null }
false
2
1
1
25
{ "stage": 2, "offload_optimizer": { "device": "none" }, "contiguous_gradients": true, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 50000000, "allgather_bucket_size": 50000000 }
{ "enabled": true }
{ "type": "AdamW", "params": { "lr": 0.00005, "betas": [ 0.9, 0.999 ], "eps": 1e-8, "weight_decay": 0 } }
{ "partition_activations": false, "cpu_checkpointing": false, "contiguous_memory_optimization": false, "number_checkpoints": null }
false
6
1
1
25
{ "stage": 2, "offload_optimizer": { "device": "none" }, "contiguous_gradients": true, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 50000000, "allgather_bucket_size": 50000000 }
{ "enabled": true }
{ "type": "AdamW", "params": { "lr": 0.00005, "betas": [ 0.9, 0.999 ], "eps": 1e-8, "weight_decay": 0 } }
{ "partition_activations": false, "cpu_checkpointing": false, "contiguous_memory_optimization": false, "number_checkpoints": null }
false
8
1
1
25
{ "stage": 2, "offload_optimizer": { "device": "none" }, "contiguous_gradients": true, "overlap_comm": true, "reduce_scatter": true, "reduce_bucket_size": 50000000, "allgather_bucket_size": 50000000 }
{ "enabled": true }
{ "type": "AdamW", "params": { "lr": 0.00005, "betas": [ 0.9, 0.999 ], "eps": 1e-8, "weight_decay": 0 } }
{ "partition_activations": false, "cpu_checkpointing": false, "contiguous_memory_optimization": false, "number_checkpoints": null }
false

GAM LIBERO Training Assets

This dataset repository contains the assets needed to fine-tune GAM on LIBERO.

Layout

checkpoints/track4world_da3.pth
data/libero_noop/<suite>/*.hdf5
data/libero_noop/_stats/*.json
configs/training/libero_unified/

track4world_da3.pth is the DA3-Giant base checkpoint. The LIBERO HDF5 files contain embedded RGB, proprioception, actions, and depth used by the public GAM training configs.

Code: https://github.com/JeonSeongHu/gam_test/tree/public/libero-da3giant Checkpoints: https://huggingface.co/SeonghuJeon/3da-libero-gam

Downloads last month
251