admin/lerobot
1
0
Fork 0
mirror of https://github.com/huggingface/lerobot.git synced 2026-05-15 16:49:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix: full pi models support for transformer v5 (#2967)

Browse Source 
* fix(pi): remove loss truncation

* fix(pi): remove state padding before tokenization

* fix(pi): fix image padding value

* fix from_pretrain

* add transformer v5 changes

* remove reference

* more fixes

* make it work

* add support for rest of pi family

* add pifast work

* more changes

* more changes

* more cleanup

* fix torch params

* dtype fix

* torch compile

* embed mismatch fix

* revert groot

* more nit fixes

* remove unused classes

* more fixes

* revert

* nit

* torch dtype warning fix

* but back dynamic renaming

* add tie embedding

---------

Co-authored-by: Yufei Sun <skieyfly@gmail.com>
This commit is contained in:
Jade Choghari
2026-02-23 22:44:13 +03:00
committed by GitHub
parent 753b996cda
commit 419305a4c2
13 changed files with 517 additions and 195 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/source/pi0fast.mdx
+10 -10
View File
@@ -52,7 +52,7 @@ This approach can transform **any existing VLM** into a VLA by training it to pr
You have two options for the FAST tokenizer:
1. **Use the pre-trained tokenizer**: The `physical-intelligence/fast` tokenizer was trained on 1M+ real robot action sequences and works as a general-purpose tokenizer.
1. **Use the pre-trained tokenizer**: The `lerobot/fast-action-tokenizer` tokenizer was trained on 1M+ real robot action sequences and works as a general-purpose tokenizer.
2. **Train your own tokenizer**: For maximum performance on your specific dataset, you can finetune the tokenizer on your own data.
@@ -114,15 +114,15 @@ lerobot-train \
### Key Training Parameters
| Parameter | Description | Default |
| -------------------------------------- | -------------------------------------------------- | ---------------------------- |
| `--policy.gradient_checkpointing=true` | Reduces memory usage significantly during training | `false` |
| `--policy.dtype=bfloat16` | Use mixed precision training for efficiency | `float32` |
| `--policy.chunk_size` | Number of action steps to predict (action horizon) | `50` |
| `--policy.n_action_steps` | Number of action steps to execute | `50` |
| `--policy.max_action_tokens` | Maximum number of FAST tokens per action chunk | `256` |
| `--policy.action_tokenizer_name` | FAST tokenizer to use | `physical-intelligence/fast` |
| `--policy.compile_model=true` | Enable torch.compile for faster training | `false` |
| Parameter | Description | Default |
| -------------------------------------- | -------------------------------------------------- | ------------------------------- |
| `--policy.gradient_checkpointing=true` | Reduces memory usage significantly during training | `false` |
| `--policy.dtype=bfloat16` | Use mixed precision training for efficiency | `float32` |
| `--policy.chunk_size` | Number of action steps to predict (action horizon) | `50` |
| `--policy.n_action_steps` | Number of action steps to execute | `50` |
| `--policy.max_action_tokens` | Maximum number of FAST tokens per action chunk | `256` |
| `--policy.action_tokenizer_name` | FAST tokenizer to use | `lerobot/fast-action-tokenizer` |
| `--policy.compile_model=true` | Enable torch.compile for faster training | `false` |
## Inference