-
Notifications
You must be signed in to change notification settings - Fork 1k
Issues
is:issue state:open
is:issue state:open
Issue creation is restricted in this repository
Search results
[Bug] Multi-head MTP (
--mtp-num-layers > 1) crashes at training-step loggingbugSomething isn't workingSomething isn't workingStatus: Open.#2131 In THUDM/slime;[Bug] When making minimax m2.7 hf checkpoint to torch_dist format, ran into error
bugSomething isn't workingSomething isn't workingStatus: Open.#2129 In THUDM/slime;- Status: Open.#2104 In THUDM/slime;
[Bug] slime-v0.3.0 版本在跑 qwen3.6 35B A3B 模型的时候,在第二次 rollout 会有乱码。怀疑 镜像&sglang 版本导致
bugSomething isn't workingSomething isn't workingStatus: Open.#2091 In THUDM/slime;[Question] Need help to support Qwen3.5 dense(/moe) VLM megatron.bridge plugin together
questionFurther information is requestedFurther information is requestedStatus: Open.#2073 In THUDM/slime;[Question] code agent rl 数据格式问题
questionFurther information is requestedFurther information is requestedStatus: Open.#2052 In THUDM/slime;[Question] torch_memory_saver 报错only hook_mode=preload supports
questionFurther information is requestedFurther information is requestedStatus: Open.#2018 In THUDM/slime;[Question] Any plans to support pipeline RL to avoid ramp down time during weight update in sglang servers
questionFurther information is requestedFurther information is requestedStatus: Open.#2007 In THUDM/slime;[Proposal] TCOD — extending slime's On-Policy Distillation to multi-turn agents
questionFurther information is requestedFurther information is requestedStatus: Open.#2002 In THUDM/slime;[RFC] Integrate TransferQueue into slime as an Optional Training Data Plane
questionFurther information is requestedFurther information is requestedStatus: Open.#1971 In THUDM/slime;[Question] retool example: compute_log_probs(logits.clone(), tokens, tp_group) torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 58.15 GiB.
questionFurther information is requestedFurther information is requestedStatus: Open.#1951 In THUDM/slime;[Bug] TorchMemorySaver observes invalid LD_PRELOAD. when add --disable-weights-backuper
bugSomething isn't workingSomething isn't workingStatus: Open.#1936 In THUDM/slime;