Reinforcement Learning
PEFT
Safetensors
grpo
trl
RL Environment
openenv
p5js
generative-art
lora
sergiopaniego HF Staff commited on
Commit
55940db
·
verified ·
1 Parent(s): 329cabb

Drop internal run names from the card

Browse files
Files changed (1) hide show
  1. README.md +2 -3
README.md CHANGED
@@ -67,7 +67,7 @@ tower, which a text-only run was never going to touch.
67
 
68
  ```bash
69
  hf jobs uv run examples/watercolour_grpo.py --flavor h200 --timeout 24h --secrets HF_TOKEN -- \
70
- --env-url https://sergiopaniego-watercolour-env-v20.hf.space \
71
  --model Qwen/Qwen3.5-35B-A3B --lora --all-linear --bf16 --gradient-checkpointing \
72
  --subject 'a peach hibiscus' --references 4 \
73
  --top-p 0.95 --top-k 20 \
@@ -130,8 +130,7 @@ generalisation probe is pending.
130
  judge. It is not cheap.
131
 
132
  Method reproduced from [Surya Narreddi's "RL'ing Qwen to paint with
133
- code"](https://surya.website/rling-qwen-to-paint-with-code). Internally this run is `v20`,
134
- HF job `6a936642984507d9db4ec2a9`.
135
 
136
  ## Where this comes from
137
 
 
67
 
68
  ```bash
69
  hf jobs uv run examples/watercolour_grpo.py --flavor h200 --timeout 24h --secrets HF_TOKEN -- \
70
+ --env-url https://<you>-watercolour-env.hf.space \
71
  --model Qwen/Qwen3.5-35B-A3B --lora --all-linear --bf16 --gradient-checkpointing \
72
  --subject 'a peach hibiscus' --references 4 \
73
  --top-p 0.95 --top-k 20 \
 
130
  judge. It is not cheap.
131
 
132
  Method reproduced from [Surya Narreddi's "RL'ing Qwen to paint with
133
+ code"](https://surya.website/rling-qwen-to-paint-with-code). Trained on HF Jobs (job `6a936642984507d9db4ec2a9`).
 
134
 
135
  ## Where this comes from
136