Workshop DQN teacher (2000000 decisions)

README.md

@@ -0,0 +1,11 @@
+# slm-rl-boxing-dqn
+
+CleanRL-pattern DQN teacher for ALE boxing (SLM-RL).
+
+- Checkpoint: dqn.pt (2000000 decisions)
+- Matching demo pack: BLANK/slm-rl-boxing
+- Matching LoRA: BLANK/slm-rl-boxing
+
+Paste this repo as the playground dqn_url when needed.
+
+Trained with [SLM-RL](https://github.com/CraftsMan-Labs/SLM-RL).

dqn.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57646e0a0869931f33da0d2a9fdbcb27a3bd10204e6f003ec8a2c6ab80c1dd1b
+size 408925