Fix tokenizer reading and update README
This commit is contained in:
+3
-3
@@ -19,10 +19,10 @@ git clone https://huggingface.co/physical-intelligence/fast
|
||||
|
||||
## Required Paths (Must Modify)
|
||||
```yaml
|
||||
pretrained_wallx_path: "/path/to/wallx_model/" # Path to pretrained Qwen VL model
|
||||
use_fast_tokenizer: false # True: train FAST, False: train Flow
|
||||
action_tokenizer_path: "/path/to/fast/" # Path to action tokenizer
|
||||
pretrained_wallx_path: "/path/to/wallx_model/" # Path to pretrained wallx model
|
||||
save_path: "/path/to/workspace/" # Path to save training outputs
|
||||
use_fast_tokenizer: False # True: train FAST, False: train Flow
|
||||
action_tokenizer_path: "/path/to/fast/" # Must set if use_fast_tokenizer is True
|
||||
```
|
||||
|
||||
## Training Parameters (Commonly Modified)
|
||||
|
||||
Reference in New Issue
Block a user