g1 open_lid_add_potato FDM-v2 transformer @ step 500
Browse files- README.md +39 -0
- transformer/config.json +23 -0
- transformer/diffusion_pytorch_model.safetensors +3 -0
README.md
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- robotics
|
| 5 |
+
- lingbot-va
|
| 6 |
+
- unitree-g1
|
| 7 |
+
- world-model
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# g1_fdmv2_openLidPotato_500step — LingBot-VA G1 post-trained transformer
|
| 11 |
+
|
| 12 |
+
Fine-tuned `transformer` for LingBot-VA on Unitree G1 (Dex1), task
|
| 13 |
+
`bobchenyx/g1_dex1_open_lid_add_potato_033`:
|
| 14 |
+
*"Open the pot's lid and put the potato inside the pot."*
|
| 15 |
+
|
| 16 |
+
- Base: `robbyant/lingbot-va-base`
|
| 17 |
+
- Post-training: 50 demos, single task, lr 1e-5, **FDM v2 recipe** —
|
| 18 |
+
mutually-exclusive per-microstep regime (rank-synced coin `fdm_prob=0.5`:
|
| 19 |
+
FDM video-only L_fdm Eq.13 `lambda_fdm=1.0` OR standard IDM L_dyn+L_inv;
|
| 20 |
+
one forward, one backward). Optimizer **step 500** of a 2000-step run
|
| 21 |
+
(training still in progress).
|
| 22 |
+
- This repo contains **only `transformer/`** — `vae/`, `text_encoder/`,
|
| 23 |
+
`tokenizer/` are unchanged from `robbyant/lingbot-va-base`.
|
| 24 |
+
|
| 25 |
+
## Assemble an eval-ready checkpoint
|
| 26 |
+
|
| 27 |
+
```bash
|
| 28 |
+
hf download robbyant/lingbot-va-base --local-dir lingbot-va-base
|
| 29 |
+
hf download armanakbari4/g1_fdmv2_openLidPotato_500step --local-dir g1_olp_500_dl
|
| 30 |
+
|
| 31 |
+
mkdir -p g1_olp_500
|
| 32 |
+
ln -sf $(realpath g1_olp_500_dl/transformer) g1_olp_500/transformer
|
| 33 |
+
ln -sf $(realpath lingbot-va-base/vae) g1_olp_500/vae
|
| 34 |
+
ln -sf $(realpath lingbot-va-base/text_encoder) g1_olp_500/text_encoder
|
| 35 |
+
ln -sf $(realpath lingbot-va-base/tokenizer) g1_olp_500/tokenizer
|
| 36 |
+
```
|
| 37 |
+
|
| 38 |
+
Serve with `CONFIG_NAME=g1_openlidpotato MODEL_PATH=g1_olp_500`.
|
| 39 |
+
`transformer/config.json` has `attn_mode: torch` (inference-ready).
|
transformer/config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"patch_size": [
|
| 3 |
+
1,
|
| 4 |
+
2,
|
| 5 |
+
2
|
| 6 |
+
],
|
| 7 |
+
"num_attention_heads": 24,
|
| 8 |
+
"attention_head_dim": 128,
|
| 9 |
+
"in_channels": 48,
|
| 10 |
+
"out_channels": 48,
|
| 11 |
+
"action_dim": 30,
|
| 12 |
+
"text_dim": 4096,
|
| 13 |
+
"freq_dim": 256,
|
| 14 |
+
"ffn_dim": 14336,
|
| 15 |
+
"num_layers": 30,
|
| 16 |
+
"cross_attn_norm": true,
|
| 17 |
+
"eps": 1e-06,
|
| 18 |
+
"rope_max_seq_len": 1024,
|
| 19 |
+
"pos_embed_seq_len": null,
|
| 20 |
+
"attn_mode": "torch",
|
| 21 |
+
"_class_name": "WanTransformer3DModel",
|
| 22 |
+
"_diffusers_version": "0.35.0.dev0"
|
| 23 |
+
}
|
transformer/diffusion_pytorch_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01800d8db63001d02bf87e22922cdaae1c6bdcb948f336cc9acd77aadaf17559
|
| 3 |
+
size 10177831668
|