|
|
@@ -73,15 +73,21 @@
|
|
|
|
|
|
engine = "vllm"
|
|
|
name = "GLM-5.3-Flash NVFP4-Spark (vLLM, 2x Spark, MTP-5)"
|
|
|
-# karmic-kraken-beta — the lab's unified GLM/DS image (2026-09-19, per the
|
|
|
-# rtx6kpro glm-5.3-flash-spark-tp2 doc): CUDA 13.4.1, B12X backends, the
|
|
|
-# NVFP4-Spark loader + MTP head support (the old r12 image is gone from the
|
|
|
-# repo as non-rebuildable, commit 17a91a9; the prebuilt lab image replaces
|
|
|
-# it). Prebuilt external image — no docker_image_dir.
|
|
|
-docker_image = "ghcr.io/local-inference-lab/vllm:karmic-kraken-beta"
|
|
|
+# Image: glm53-jovian-r35:local — our r35 anchor image for the GLM-5.3-Flash
|
|
|
+# NVFP4 family (the 433 GB 4x lane ran on it at 163/176 with MTP3 — the
|
|
|
+# Spark checkpoint is the same Glm5Next family, different quant variant).
|
|
|
+# The lab's newer karmic-kraken-beta (rtx6kpro spark-tp2 doc) is amd64-only
|
|
|
+# (their RTX 6000 x86 platform; "Exec format error" on our aarch64 nodes),
|
|
|
+# so it cannot serve the Spark checkpoint here; the r12 image (the recipe's
|
|
|
+# original) is gone from the repo as non-rebuildable (commit 17a91a9).
|
|
|
+# Aiden's production GLM-5.3-Flash image for GB10 (31.2 GB; the glm53-flash-aiden
|
|
|
+# lane ran on it at 100/100 agentic — the GDN+NVFP4 path routes through B12X,
|
|
|
+# which handles the Spark checkpoint's small KDA GEMMs that r35's flashinfer
|
|
|
+# mm_mxfp8 (n,k >= 128) rejects. Prebuilt external image — no image_dir.
|
|
|
+docker_image = "aidendle94/sparkrun-vllm-glm53-gb10:production-1.2"
|
|
|
# Required: the clean image follows the lab contract (ENTRYPOINT []
|
|
|
# + CMD bash), so the recipe carries the entrypoint.
|
|
|
-entrypoint = ["vllm", "serve"]
|
|
|
+entrypoint = []
|
|
|
model = "local-inference-lab/GLM-5.3-Flash-NVFP4-Spark"
|
|
|
parent_model = "zai-org/GLM-5.3-Flash"
|
|
|
required_nodes = 2
|
