Update README.md
Browse files
README.md
CHANGED
|
@@ -59,4 +59,8 @@ Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a
|
|
| 59 |
|
| 60 |
## Model Details
|
| 61 |
|
| 62 |
-
- **License:** apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
## Model Details
|
| 61 |
|
| 62 |
+
- **License:** apache-2.0
|
| 63 |
+
|
| 64 |
+
## Comments
|
| 65 |
+
- First trained with 40 episodes for ~3 epochs. The arm was moving decently well, but was unable to come down near the table level (as shown in demo) in the grabbing pose.
|
| 66 |
+
- Then collected 60 episodes more, and trained for ~8 epochs from scratch. Didn't see any difference between each of the policy, even though trained for more epoch, and more episodes. Had the same issue in coming down.
|