sglang-glm53-flash-sm120
SGLang GLM-5.3-Flash on 2x RTX PRO 6000 Blackwell (SM120, TP2): C4 @ 507k-token KV pool, W4A16+FP8-mix quant, HiCache, native MTP — reproducible image, producers, bench harness, receipts
File Explorer
- build-release.yml
- build-test.yml
- package-retention.yml
- promote-release.yml
- validate.yml
- build-image.yml
- promote-release.yml
- validate.yml
- buildkitd.toml
- acceptance_check.cpython-314.pyc
- agentx-mvp.yaml
- decode-engine.yaml
- decode-steady.yaml
- gsm8k.yaml
- prefill-cold-vllm.yaml
- prefill-cold.yaml
- turnover.yaml
- warmup-coverage.yaml
- aiperf.lock.json
- analyze_pairs.py
- analyze_prefill.py
- analyze_server_metrics.py
- analyze_turnover.py
- analyze_vllm_server_metrics.py
- assemble_observations.py
- block_manifest.py
- build_block_results.py
- CALIBRATION-2026-08-10.md
- capture-environment.sh
- compare_engine_gates.py
- compare_turnover_gates.py
- decode-engine.yaml
- prepare-in-pod.sh
- README.md
- run-agentx-gate-in-pod.sh
- run-engine-gate-in-pod.sh
- run-in-pod.sh
- run-performance-block-in-pod.sh
- run-turnover-gate-in-pod.sh
- seed-panel.json
- stage-uv-in-pod.sh
- STATISTICAL-DESIGN.md
- summarize_engine_gate.py
- summarize_turnover_gate.py
- test_analyze_pairs.py
- test_analyze_prefill.py
- test_analyze_server_metrics.py
- test_analyze_turnover.py
- test_analyze_vllm_server_metrics.py
- test_assemble_observations.py
- test_block_manifest.py
- test_build_block_results.py
- test_compare_engine_gates.py
- test_compare_turnover_gates.py
- test_summarize_engine_gate.py
- test_summarize_turnover_gate.py
- qualify_openai.py
- regrade_gsm8k.py
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- decode-analysis.json
- environment.txt
- benchmark-config.yaml
- environment.txt
- prefill-analysis.json
- benchmark-config.yaml
- environment.txt
- prefill-analysis.json
- benchmark-config.yaml
- environment.txt
- prefill-analysis.json
- benchmark-config.yaml
- environment.txt
- prefill-analysis.json
- summary.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- completed-at-utc.txt
- SHA256SUMS
- summary.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- completed-at-utc.txt
- SHA256SUMS
- summary.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- decode-analysis.json
- completed-at-utc.txt
- SHA256SUMS
- summary.json
- v0.2.0-rc.1-validation-20260903.txt
- v020rc1-c1-fusionoff-summary.md
- v020rc1-c4-summary.md
- v020rc1-qual-summary.md
- v0.1.0-rc.10-build-failure-20260828.txt
- v0.1.0-rc.64-adaptive-ladder-promotion-20260831.txt
- v0.1.0-rc.64-c1-decode-artifact-ab-20260831.txt
- v0.1.0-rc.64-c2-c4-decode-cells-20260831.txt
- v0.1.0-rc.64-decode-plateaus-corrected-20260831.txt
- v0.1.0-rc.64-nvfp4-experts-fp8-wo-release-sweep-20260831.txt
- v0.1.0-rc.64-transport-ab-20260831.txt
- v0.1.0-rc.65-dflash2-evaluation-20260831.txt
- v0.1.1-capacity-ladder-20260902.txt
- v0.1.1-capacity-ladder-job-20260902.log
- v0.1.1-engine-gate-decode-prefill-20260902.txt
- v0.1.1-gsm8k-full-quality-20260902.txt
- v0.1.1-rc.16-gsm8k-100-20260902.json
- v0.1.2-rc.1-c4-admission-20260902.txt
- v0.1.2-rc.1-gsm8k-100-20260902.json
- v0.1.3-headroom-study-20260902.txt
- v0.1.4-padded-replay-recheck-20260902.txt
- v0.1.4-rc.1-validation-20260902.txt
- v0.1.4-rc.2-validation-20260902.txt
- glm53-adaptive.json
- serve-glm53-flash.sh
- flashinfer-glm53-integration.patch
- sglang-glm53-integration.patch
- quantize_glm53_bf16_w4a16_e4m3_k32.py
- quantize_glm53_bf16_w4a16_k32_fp8attn_mix.py
- prune-container-releases.sh
- validate-docs.sh
- validate-release.sh
- verify-patches.sh
- .gitattributes
- acceptance_check.py
- AGENTS.md
- BENCHMARKS.md
- CHANGELOG.md
- Containerfile
- LICENSE
- NOTICE.md
- QUANTIZATION.md
- README.md
- release.json
- RUN.md
- stack.lock.json
# Use via CDN
jsDelivrjsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.
Link
Example
// repository documentation
Was this content helpful?
(0 ratings)