restore: bring back deleted examples, scripts, and agent doc - #220
Closed
aoshen02 wants to merge 2 commits into
Closed
restore: bring back deleted examples, scripts, and agent doc#220aoshen02 wants to merge 2 commits into
aoshen02 wants to merge 2 commits into
Conversation
aoshen02
force-pushed
the
restore-examples
branch
2 times, most recently
from
June 9, 2026 15:23
5ec2e63 to
880235a
Compare
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive roadmap documentation for Agentic RL training alongside several bash scripts to run various training configurations (such as fully async, low-precision FP8, and SFT) for Qwen, GLM, and Moonlight models. The review feedback highlights several critical issues: incorrect relative source paths in the low-precision scripts, unescaped double quotes in an SSH command that would break remote execution, broken documentation links pointing to non-existent directories, and the repeated use of the unrecognized environment variable PYTHONBUFFERED instead of PYTHONUNBUFFERED=1 across multiple scripts.
| echo "HAS_NVLINK: $HAS_NVLINK (detected $NVLINK_COUNT NVLink references)" | ||
|
|
||
| SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)" | ||
| source "${SCRIPT_DIR}/../scripts/models/qwen3-30B-A3B.sh" |
Contributor
There was a problem hiding this comment.
The source path is incorrect. It resolves to scripts/scripts/models/qwen3-30B-A3B.sh because SCRIPT_DIR is scripts/low_precision. It should be ../models/qwen3-30B-A3B.sh instead of ../scripts/models/qwen3-30B-A3B.sh.
Suggested change
| source "${SCRIPT_DIR}/../scripts/models/qwen3-30B-A3B.sh" | |
| source "${SCRIPT_DIR}/../models/qwen3-30B-A3B.sh" |
| echo "HAS_NVLINK: $HAS_NVLINK (detected $NVLINK_COUNT NVLink references)" | ||
|
|
||
| SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &>/dev/null && pwd)" | ||
| source "${SCRIPT_DIR}/../scripts/models/qwen3-4B.sh" |
Contributor
There was a problem hiding this comment.
The source path is incorrect. It resolves to scripts/scripts/models/qwen3-4B.sh because SCRIPT_DIR is scripts/low_precision. It should be ../models/qwen3-4B.sh instead of ../scripts/models/qwen3-4B.sh.
Suggested change
| source "${SCRIPT_DIR}/../scripts/models/qwen3-4B.sh" | |
| source "${SCRIPT_DIR}/../models/qwen3-4B.sh" |
Comment on lines
+132
to
+133
| ssh root@"${WORKER_IP}" \ | ||
| "pkill -9 -f "vllm serve" ; ray stop --force ; pkill -9 python ; ray start --address=${MASTER_ADDR}:6379 --num-gpus 8 --node-ip-address ${WORKER_IP} --disable-usage-stats --dashboard-host=0.0.0.0 --dashboard-port=8265" & |
Contributor
There was a problem hiding this comment.
The inner double quotes around vllm serve are not escaped, which causes them to be stripped by the local shell before executing the ssh command. This results in executing pkill -9 -f vllm serve on the remote worker, which is incorrect. Escape the inner double quotes to preserve them.
Suggested change
| ssh root@"${WORKER_IP}" \ | |
| "pkill -9 -f "vllm serve" ; ray stop --force ; pkill -9 python ; ray start --address=${MASTER_ADDR}:6379 --num-gpus 8 --node-ip-address ${WORKER_IP} --disable-usage-stats --dashboard-host=0.0.0.0 --dashboard-port=8265" & | |
| ssh root@"${WORKER_IP}" \ | |
| "pkill -9 -f \"vllm serve\" ; ray stop --force ; pkill -9 python ; ray start --address=${MASTER_ADDR}:6379 --num-gpus 8 --node-ip-address ${WORKER_IP} --disable-usage-stats --dashboard-host=0.0.0.0 --dashboard-port=8265" & |
Comment on lines
+13
to
+15
| | Return multiple training samples from one prompt, such as subagent, multi-agent, or context-compaction segments | [fan-out return from custom generate](customization.md#returning-multiple-training-samples-for-one-prompt), [`examples/multi_agent`](../_examples_synced/multi_agent/README.md) | | ||
| | Avoid blocking training on long-tail agent rollouts | [`examples/fully_async`](../_examples_synced/fully_async/README.md) | | ||
| | Study a full end-to-end agent example with sandboxing, real code edits, and test-based grading | [`examples/coding_agent_rl`](../_examples_synced/coding_agent_rl/README.md) | |
Contributor
There was a problem hiding this comment.
The links point to ../_examples_synced/... which does not exist in this repository. They should be updated to point to the actual examples directory relative to this file (i.e., ../../../examples/...).
Suggested change
| | Return multiple training samples from one prompt, such as subagent, multi-agent, or context-compaction segments | [fan-out return from custom generate](customization.md#returning-multiple-training-samples-for-one-prompt), [`examples/multi_agent`](../_examples_synced/multi_agent/README.md) | | |
| | Avoid blocking training on long-tail agent rollouts | [`examples/fully_async`](../_examples_synced/fully_async/README.md) | | |
| | Study a full end-to-end agent example with sandboxing, real code edits, and test-based grading | [`examples/coding_agent_rl`](../_examples_synced/coding_agent_rl/README.md) | | |
| | Return multiple training samples from one prompt, such as subagent, multi-agent, or context-compaction segments | [fan-out return from custom generate](customization.md#returning-multiple-training-samples-for-one-prompt), [`examples/multi_agent`](../../../examples/multi_agent/README.md) | | |
| | Avoid blocking training on long-tail agent rollouts | [`examples/fully_async`](../../../examples/fully_async/README.md) | | |
| | Study a full end-to-end agent example with sandboxing, real code edits, and test-based grading | [`examples/coding_agent_rl`](../../../examples/coding_agent_rl/README.md) | |
Comment on lines
+69
to
+73
| The full coding-agent example is [`examples/coding_agent_rl`](../_examples_synced/coding_agent_rl/README.md). It shows an end-to-end agent RL setup that is close to a real software-engineering workflow: each sample boots an isolated sandbox, the agent uses tools to edit code, the rollout captures a `git diff`, and a clean sandbox runs the tests to produce the reward. | ||
|
|
||
| This example also demonstrates agent fan-out training. Its middleware splits one trajectory into `subagent`, `wipe` (the chain frozen before compaction), and `final` segments. `generate()` returns `list[Sample]`, and all segments share the same `rollout_id`. | ||
|
|
||
| For smaller starting points, see [`examples/search-r1`](../_examples_synced/search-r1/README.md) for multi-turn tool use, [`examples/retool`](../_examples_synced/retool/README.md) for tool-augmented generation, and [`examples/multi_agent`](../_examples_synced/multi_agent/README.md) for the multi-agent pattern. |
Contributor
There was a problem hiding this comment.
The links point to ../_examples_synced/... which does not exist in this repository. They should be updated to point to the actual examples directory relative to this file (i.e., ../../../examples/...).
Suggested change
| The full coding-agent example is [`examples/coding_agent_rl`](../_examples_synced/coding_agent_rl/README.md). It shows an end-to-end agent RL setup that is close to a real software-engineering workflow: each sample boots an isolated sandbox, the agent uses tools to edit code, the rollout captures a `git diff`, and a clean sandbox runs the tests to produce the reward. | |
| This example also demonstrates agent fan-out training. Its middleware splits one trajectory into `subagent`, `wipe` (the chain frozen before compaction), and `final` segments. `generate()` returns `list[Sample]`, and all segments share the same `rollout_id`. | |
| For smaller starting points, see [`examples/search-r1`](../_examples_synced/search-r1/README.md) for multi-turn tool use, [`examples/retool`](../_examples_synced/retool/README.md) for tool-augmented generation, and [`examples/multi_agent`](../_examples_synced/multi_agent/README.md) for the multi-agent pattern. | |
| The full coding-agent example is [`examples/coding_agent_rl`](../../../examples/coding_agent_rl/README.md). It shows an end-to-end agent RL setup that is close to a real software-engineering workflow: each sample boots an isolated sandbox, the agent uses tools to edit code, the rollout captures a `git diff`, and a clean sandbox runs the tests to produce the reward. | |
| This example also demonstrates agent fan-out training. Its middleware splits one trajectory into `subagent`, `wipe` (the chain frozen before compaction), and `final` segments. `generate()` returns `list[Sample]`, and all segments share the same `rollout_id`. | |
| For smaller starting points, see [`examples/search-r1`](../../../examples/search-r1/README.md) for multi-turn tool use, [`examples/retool`](../../../examples/retool/README.md) for tool-augmented generation, and [`examples/multi_agent`](../../../examples/multi_agent/README.md) for the multi-agent pattern. |
| set -ex | ||
|
|
||
| # will prevent ray from buffering stdout/stderr | ||
| export PYTHONUNBUFFERED=1 |
Contributor
There was a problem hiding this comment.
The environment variable PYTHONBUFFERED is not recognized by Python. The correct variable to disable stdout/stderr buffering is PYTHONUNBUFFERED=1.
Suggested change
| export PYTHONUNBUFFERED=1 | |
| export PYTHONUNBUFFERED=1 |
| set -ex | ||
|
|
||
| # will prevent ray from buffering stdout/stderr | ||
| export PYTHONUNBUFFERED=1 |
Contributor
There was a problem hiding this comment.
The environment variable PYTHONBUFFERED is not recognized by Python. The correct variable to disable stdout/stderr buffering is PYTHONUNBUFFERED=1.
Suggested change
| export PYTHONUNBUFFERED=1 | |
| export PYTHONUNBUFFERED=1 |
| set -ex | ||
|
|
||
| # will prevent ray from buffering stdout/stderr | ||
| export PYTHONUNBUFFERED=1 |
Contributor
There was a problem hiding this comment.
The environment variable PYTHONBUFFERED is not recognized by Python. The correct variable to disable stdout/stderr buffering is PYTHONUNBUFFERED=1.
Suggested change
| export PYTHONUNBUFFERED=1 | |
| export PYTHONUNBUFFERED=1 |
| set -ex | ||
|
|
||
| # will prevent ray from buffering stdout/stderr | ||
| export PYTHONUNBUFFERED=1 |
Contributor
There was a problem hiding this comment.
The environment variable PYTHONBUFFERED is not recognized by Python. The correct variable to disable stdout/stderr buffering is PYTHONUNBUFFERED=1.
Suggested change
| export PYTHONUNBUFFERED=1 | |
| export PYTHONUNBUFFERED=1 |
| # export MASTER_ADDR="127.0.0.1" | ||
|
|
||
| # will prevent ray from buffering stdout/stderr | ||
| export PYTHONUNBUFFERED=1 |
Contributor
There was a problem hiding this comment.
The environment variable PYTHONBUFFERED is not recognized by Python. The correct variable to disable stdout/stderr buffering is PYTHONUNBUFFERED=1.
Suggested change
| export PYTHONUNBUFFERED=1 | |
| export PYTHONUNBUFFERED=1 |
aoshen02
force-pushed
the
restore-examples
branch
3 times, most recently
from
June 9, 2026 15:36
9b036b1 to
a9ca353
Compare
Restore files that were either deleted by #126 ("trim examples to qwen3 only") or never synced from slime: **Reverted from pre-#126 (translated):** - scripts/low_precision/run-qwen3-4b-fp8.sh - scripts/low_precision/run-qwen3-30b-a3b-fp8.sh - scripts/run-glm4-9B.sh - scripts/run-moonlight-16B-A3B.sh - scripts/run-qwen3-4B-base-sft.sh - scripts/run-qwen3-32B.sh - scripts/run-qwen3.5-35B-A3B-sft.sh **New from slime@44d29ee (translated):** - docs/en/get_started/agent.md - examples/fully_async/run-qwen2.5-0.5B-fully_async.sh All sglang engine flags translated to vllm equivalents (§2.4). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
aoshen02
force-pushed
the
restore-examples
branch
from
June 9, 2026 15:38
a9ca353 to
f9b1e10
Compare
Standardize all scripts to use the bracket-escaped pkill pattern that avoids matching pkill itself and also catches vLLM's renamed subprocesses (VLLM::EngineCore, VLLM::Worker_TP*). Matches the canonical pattern in command_utils.py. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
aoshen02
marked this pull request as draft
June 9, 2026 17:01
aoshen02
added a commit
that referenced
this pull request
Jun 21, 2026
pkill -9 vllm matches any process named "vllm" and can inadvertently kill unrelated vllm processes (e.g. background services). Use the same pattern as PR #220 which targets only vllm serve and Ray VLL[M]:: actors: pkill -9 -f '[v]llm serve|VLL[M]::' Also updates the inline form used in multi-node SSH worker restart commands (run-qwen3-235B-A22B*.sh, run-qwen3.5-27B.sh, etc.). Skipped: scripts/run-gpt-oss-20B.sh (uses pkill -9 -f "vllm serve" already), scripts/run-minimax-m2.sh and run-glm4.7-*.sh (already used -f "vllm serve"). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
CalvinXKY
pushed a commit
that referenced
this pull request
Jun 25, 2026
… 20B support (#260) * restore: bring back deleted examples, scripts, and agent doc Restore files that were either deleted by #126 ("trim examples to qwen3 only") or never synced from slime: **Reverted from pre-#126 (translated):** - scripts/low_precision/run-qwen3-4b-fp8.sh - scripts/low_precision/run-qwen3-30b-a3b-fp8.sh - scripts/run-glm4-9B.sh - scripts/run-moonlight-16B-A3B.sh - scripts/run-qwen3-4B-base-sft.sh - scripts/run-qwen3-32B.sh - scripts/run-qwen3.5-35B-A3B-sft.sh **New from slime@44d29ee (translated):** - docs/en/get_started/agent.md - examples/fully_async/run-qwen2.5-0.5B-fully_async.sh All sglang engine flags translated to vllm equivalents (§2.4). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: unify pkill pattern to '[v]llm serve|VLL[M]::' Standardize all scripts to use the bracket-escaped pkill pattern that avoids matching pkill itself and also catches vLLM's renamed subprocesses (VLLM::EngineCore, VLLM::Worker_TP*). Matches the canonical pattern in command_utils.py. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * scripts: complete slime-exact translation of all 29 run scripts Translate all slime scripts to vime following SGLANG_TO_VLLM_TRANSLATION.md: - sglang→vllm prefix swap for CLI flags and variables - _slime→_vime for checkpoint paths - EP: --sglang-ep-size N → --vllm-enable-expert-parallel (boolean) - Speculative: multi-param → --vllm-speculative-config JSON (§5.2) - Delete genuinely sglang-coupled params (DP-attention, DeepEP, NSA, etc.) - flashinfer → FLASHINFER case fix (§2.4) 23 new scripts + 6 existing updated to match slime@cutoff. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(scripts): correct model config source path in FP8 low_precision scripts The FP8 scripts used `${SCRIPT_DIR}/../scripts/models/` which resolves to `scripts/scripts/models/` (non-existent). Changed to `../models/` to match the INT4 scripts. Same fix as slime PR #2094. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(gpt-oss): fused BF16 format, bridge API patch, bshd qkv format Three fixes needed to run GPT-OSS 20B RLHF on vLLM backend: 1. hf_weight_iterator_bridge: match Megatron-Bridge 0.5.0 API _patch_bridge_expert_cache_to_cpu monkey-patches GPTOSSBridge. maybe_modify_converted_hf_weight gained a 4th `hf_state_dict` parameter; the patched wrapper only accepted 3, causing TypeError during weight sync. 2. run-gpt-oss-20B: point --hf-checkpoint at fused BF16 format vLLM's _load_weights_other expects gate_up_proj [E, hidden, 2*ffn] (fused). The old per-expert split format (experts.{e}.gate_proj.weight) causes KeyError on bias loading. Use tools/convert_gpt_oss_to_fused.py to convert an existing per-expert checkpoint, or re-run preprocess_gpt_oss.py to produce fused format directly. 3. run-gpt-oss-20B: add --qkv-format bshd + fix seq-length GPT-OSS uses learnable softmax (sink attention). TransformerEngine disables all attention backends when softmax_type=learnable and qkv_format=thd (packed sequences). --qkv-format bshd avoids this. --use-dynamic-batch-size is incompatible with bshd; replaced with fixed --seq-length 10240 (covers 8192 max response + prompt headroom). tools/convert_gpt_oss_to_fused.py: new tool to convert per-expert BF16 checkpoint (output of old preprocess_gpt_oss.py) to the fused HF format expected by vLLM without re-running the slow MXFP4 dequantization. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix(scripts): replace pkill -9 vllm with precise -f pattern (21 files) pkill -9 vllm matches any process named "vllm" and can inadvertently kill unrelated vllm processes (e.g. background services). Use the same pattern as PR #220 which targets only vllm serve and Ray VLL[M]:: actors: pkill -9 -f '[v]llm serve|VLL[M]::' Also updates the inline form used in multi-node SSH worker restart commands (run-qwen3-235B-A22B*.sh, run-qwen3.5-27B.sh, etc.). Skipped: scripts/run-gpt-oss-20B.sh (uses pkill -9 -f "vllm serve" already), scripts/run-minimax-m2.sh and run-glm4.7-*.sh (already used -f "vllm serve"). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore(scripts): remove run-qwen3-4B-amd.sh from this PR AMD-specific script is out of scope for the gb300-complete-port PR. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * sync(docs+scripts): port docs/examples from slime-44d29ee, fix script translations - Add missing EN/ZH docs: low-precision, on-policy-distillation, get_started/agent, pd-disaggregation (heterogeneous server groups fix), examples zh docs - Add missing examples: on_policy_distillation, eval_multi_task, delta_weight_sync, geo3k images - Fix vLLM flag translations across all example docs: - --vllm-mem-fraction-static → --vllm-gpu-memory-utilization - Remove non-existent dp-attention flags (--vllm-enable-dp-attention, --vllm-dp-size, --vllm-moe-dense-tp-size, --vllm-enable-dp-lm-head, --vllm-ep-size) - --vllm-ep-num-redundant-experts → --vllm-eplb-config - --vllm-cuda-graph-bs → --vllm-max-cudagraph-capture-size - sglang speculative flags → --vllm-speculative-config JSON - GLM-4.7 MTP: method=eagle → method=mtp, num_speculative_tokens=4 → 3 - sgl-router → vllm-router; THUDM/vime → vllm-project/vime - Fix scripts: restore run-kimi-k2-Instruct/Thinking/qwen3-4B/qwen3-235B-A22B to slime-44d29ee-as-vime + pkill precision fix only; restore int4 python3 path Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert(scripts): pkill -9 -f pattern back to pkill -9 vllm, align with slime Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert(pkill): align all remaining kill patterns with slime (pkill -9 vllm) Covers examples/, docs/, tests/, and vime/utils -- previously missed in the scripts/ revert. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove gpt-oss-20B script and convert tool (moved to separate PR) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
aoshen02
added a commit
to aoshen02/vime
that referenced
this pull request
Jul 15, 2026
… 20B support (vllm-project#260) * restore: bring back deleted examples, scripts, and agent doc Restore files that were either deleted by vllm-project#126 ("trim examples to qwen3 only") or never synced from slime: **Reverted from pre-vllm-project#126 (translated):** - scripts/low_precision/run-qwen3-4b-fp8.sh - scripts/low_precision/run-qwen3-30b-a3b-fp8.sh - scripts/run-glm4-9B.sh - scripts/run-moonlight-16B-A3B.sh - scripts/run-qwen3-4B-base-sft.sh - scripts/run-qwen3-32B.sh - scripts/run-qwen3.5-35B-A3B-sft.sh **New from slime@44d29ee (translated):** - docs/en/get_started/agent.md - examples/fully_async/run-qwen2.5-0.5B-fully_async.sh All sglang engine flags translated to vllm equivalents (§2.4). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * chore: unify pkill pattern to '[v]llm serve|VLL[M]::' Standardize all scripts to use the bracket-escaped pkill pattern that avoids matching pkill itself and also catches vLLM's renamed subprocesses (VLLM::EngineCore, VLLM::Worker_TP*). Matches the canonical pattern in command_utils.py. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * scripts: complete slime-exact translation of all 29 run scripts Translate all slime scripts to vime following SGLANG_TO_VLLM_TRANSLATION.md: - sglang→vllm prefix swap for CLI flags and variables - _slime→_vime for checkpoint paths - EP: --sglang-ep-size N → --vllm-enable-expert-parallel (boolean) - Speculative: multi-param → --vllm-speculative-config JSON (§5.2) - Delete genuinely sglang-coupled params (DP-attention, DeepEP, NSA, etc.) - flashinfer → FLASHINFER case fix (§2.4) 23 new scripts + 6 existing updated to match slime@cutoff. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(scripts): correct model config source path in FP8 low_precision scripts The FP8 scripts used `${SCRIPT_DIR}/../scripts/models/` which resolves to `scripts/scripts/models/` (non-existent). Changed to `../models/` to match the INT4 scripts. Same fix as slime PR #2094. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * fix(gpt-oss): fused BF16 format, bridge API patch, bshd qkv format Three fixes needed to run GPT-OSS 20B RLHF on vLLM backend: 1. hf_weight_iterator_bridge: match Megatron-Bridge 0.5.0 API _patch_bridge_expert_cache_to_cpu monkey-patches GPTOSSBridge. maybe_modify_converted_hf_weight gained a 4th `hf_state_dict` parameter; the patched wrapper only accepted 3, causing TypeError during weight sync. 2. run-gpt-oss-20B: point --hf-checkpoint at fused BF16 format vLLM's _load_weights_other expects gate_up_proj [E, hidden, 2*ffn] (fused). The old per-expert split format (experts.{e}.gate_proj.weight) causes KeyError on bias loading. Use tools/convert_gpt_oss_to_fused.py to convert an existing per-expert checkpoint, or re-run preprocess_gpt_oss.py to produce fused format directly. 3. run-gpt-oss-20B: add --qkv-format bshd + fix seq-length GPT-OSS uses learnable softmax (sink attention). TransformerEngine disables all attention backends when softmax_type=learnable and qkv_format=thd (packed sequences). --qkv-format bshd avoids this. --use-dynamic-batch-size is incompatible with bshd; replaced with fixed --seq-length 10240 (covers 8192 max response + prompt headroom). tools/convert_gpt_oss_to_fused.py: new tool to convert per-expert BF16 checkpoint (output of old preprocess_gpt_oss.py) to the fused HF format expected by vLLM without re-running the slow MXFP4 dequantization. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * fix(scripts): replace pkill -9 vllm with precise -f pattern (21 files) pkill -9 vllm matches any process named "vllm" and can inadvertently kill unrelated vllm processes (e.g. background services). Use the same pattern as PR vllm-project#220 which targets only vllm serve and Ray VLL[M]:: actors: pkill -9 -f '[v]llm serve|VLL[M]::' Also updates the inline form used in multi-node SSH worker restart commands (run-qwen3-235B-A22B*.sh, run-qwen3.5-27B.sh, etc.). Skipped: scripts/run-gpt-oss-20B.sh (uses pkill -9 -f "vllm serve" already), scripts/run-minimax-m2.sh and run-glm4.7-*.sh (already used -f "vllm serve"). Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore(scripts): remove run-qwen3-4B-amd.sh from this PR AMD-specific script is out of scope for the gb300-complete-port PR. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * sync(docs+scripts): port docs/examples from slime-44d29ee, fix script translations - Add missing EN/ZH docs: low-precision, on-policy-distillation, get_started/agent, pd-disaggregation (heterogeneous server groups fix), examples zh docs - Add missing examples: on_policy_distillation, eval_multi_task, delta_weight_sync, geo3k images - Fix vLLM flag translations across all example docs: - --vllm-mem-fraction-static → --vllm-gpu-memory-utilization - Remove non-existent dp-attention flags (--vllm-enable-dp-attention, --vllm-dp-size, --vllm-moe-dense-tp-size, --vllm-enable-dp-lm-head, --vllm-ep-size) - --vllm-ep-num-redundant-experts → --vllm-eplb-config - --vllm-cuda-graph-bs → --vllm-max-cudagraph-capture-size - sglang speculative flags → --vllm-speculative-config JSON - GLM-4.7 MTP: method=eagle → method=mtp, num_speculative_tokens=4 → 3 - sgl-router → vllm-router; THUDM/vime → vllm-project/vime - Fix scripts: restore run-kimi-k2-Instruct/Thinking/qwen3-4B/qwen3-235B-A22B to slime-44d29ee-as-vime + pkill precision fix only; restore int4 python3 path Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert(scripts): pkill -9 -f pattern back to pkill -9 vllm, align with slime Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * revert(pkill): align all remaining kill patterns with slime (pkill -9 vllm) Covers examples/, docs/, tests/, and vime/utils -- previously missed in the scripts/ revert. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * chore: remove gpt-oss-20B script and convert tool (moved to separate PR) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Signed-off-by: aoshen02 <aoshen@inferact.ai>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Restore 9 files that were either deleted by #126 ("trim examples to qwen3 only") or never synced from slime:
Reverted from pre-#126:
scripts/low_precision/run-qwen3-4b-fp8.shscripts/low_precision/run-qwen3-30b-a3b-fp8.shscripts/run-glm4-9B.shscripts/run-moonlight-16B-A3B.shscripts/run-qwen3-4B-base-sft.shscripts/run-qwen3-32B.shscripts/run-qwen3.5-35B-A3B-sft.shNew from slime@44d29ee (never on main):
docs/en/get_started/agent.mdexamples/fully_async/run-qwen2.5-0.5B-fully_async.shAll sglang engine flags translated to vllm equivalents per §2.4.
Test plan
--vllm-*flagsagent.mdlinks and references are valid🤖 Generated with Claude Code