Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -32,6 +32,13 @@ _This is a **deliberately published failure** — a documented negative result.
|
|
| 32 |
- [vitorcen/LeIsaac-Training](https://github.com/vitorcen/LeIsaac-Training) — LeIsaac fork（训练脚本 + 设计文档 / training scripts + design docs）
|
| 33 |
- 完整 negative report HTML: [`pi05_pytorch_expert_ft_negative.html`](https://github.com/vitorcen/LeIsaac-Training/blob/main/docs/training/pi05_pytorch_expert_ft_negative.html)
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
## TL;DR
|
| 36 |
|
| 37 |
| Item | Value |
|
|
@@ -134,11 +141,14 @@ policy = PI05Policy.from_pretrained("wsagi/Pi0.5-PickOrange")
|
|
| 134 |
|
| 135 |
## 替代方案 / Alternatives (better on same task)
|
| 136 |
|
|
|
|
|
|
|
| 137 |
| Model | Strict | Where |
|
| 138 |
|-------|--------|-------|
|
| 139 |
-
| 🥇 GR00T-N1.7 (self-trained) | 68.3% | [`wsagi/GR00T-N1.
|
| 140 |
-
| 🥈
|
| 141 |
-
| 🥉
|
|
|
|
| 142 |
|
| 143 |
## License & Attribution
|
| 144 |
|
|
|
|
| 32 |
- [vitorcen/LeIsaac-Training](https://github.com/vitorcen/LeIsaac-Training) — LeIsaac fork（训练脚本 + 设计文档 / training scripts + design docs）
|
| 33 |
- 完整 negative report HTML: [`pi05_pytorch_expert_ft_negative.html`](https://github.com/vitorcen/LeIsaac-Training/blob/main/docs/training/pi05_pytorch_expert_ft_negative.html)
|
| 34 |
|
| 35 |
+
## 🎥 失败现场录屏 / The failure, on video
|
| 36 |
+
|
| 37 |
+
<video controls src="https://huggingface.co/wsagi/Pi0.5-PickOrange/resolve/main/Pi0.5-PickOrange.mp4"></video>
|
| 38 |
+
|
| 39 |
+
_π0.5 expert-FT ckpt 在 LeIsaac PickOrange 上的真实录屏：机械臂持续运动满 180s，橙子一颗未入盘（**0/3**）。这不是 bug，是 SigLIP@224 vision bottleneck 下"看不见橙子"的真实表现——和成功模型（GR00T-N1.7 / ACT）形成直接对照。_
|
| 40 |
+
_Real screen capture: the arm keeps moving for the full 180s but places **0/3** oranges. Not a bug — the genuine behavior under the SigLIP@224 vision bottleneck. Compare against the models that actually succeed (GR00T-N1.7 / ACT) below._
|
| 41 |
+
|
| 42 |
## TL;DR
|
| 43 |
|
| 44 |
| Item | Value |
|
|
|
|
| 141 |
|
| 142 |
## 替代方案 / Alternatives (better on same task)
|
| 143 |
|
| 144 |
+
这些是**同任务上真能把橙子夹进盘子**的模型 — 想看成功的就去这里 / models that actually place the orange:
|
| 145 |
+
|
| 146 |
| Model | Strict | Where |
|
| 147 |
|-------|--------|-------|
|
| 148 |
+
| 🥇 GR00T-N1.7 (self-trained) | **68.3%** (2.05/3) | [`wsagi/GR00T-N1.7-PickOrange`](https://huggingface.co/wsagi/GR00T-N1.7-PickOrange) |
|
| 149 |
+
| 🥈 ACT (self, h=70) | **43.3%** (1.30/3) | [`wsagi/ACT-PickOrange`](https://huggingface.co/wsagi/ACT-PickOrange) |
|
| 150 |
+
| 🥉 SmolVLA (self-trained) | 25.0% | wsagi (待发布 / pending) |
|
| 151 |
+
| Diffusion Policy DDIM | 概率性 3/3 / stochastic | [`wsagi/DiffusionPolicy-PickOrange`](https://huggingface.co/wsagi/DiffusionPolicy-PickOrange) |
|
| 152 |
|
| 153 |
## License & Attribution
|
| 154 |
|