manifest Qwen/Qwen3.5-0.8B @ B300 (5f91337d78febad4)

Qwen__Qwen3.5-0.8B/B300/tp1-seq131072-lora64x4/5f91337d78febad4/manifest.json

@@ -38,7 +38,7 @@
   "image_tag": "baseten/baseten-weight-sync-inference:main-68cee18",
   "caller": "github-actions:William-Gao1",
   "model_revision": "2fc06364715b967f1860aea9cf38778875588b17",
-  "build_id": "29971523594-1",
+  "build_id": "29972324745-1",
   "opaque_sampler_payload": {
     "tensor_parallel_size": 1,
     "max_seq_length": 131072,
@@ -94,13 +94,13 @@
     "use_mega_aot_artifact": true,
     "deep_gemm_warmup": "skip"
   },
-  "ready_in_seconds_no_cache": 400.15,
+  "ready_in_seconds_no_cache": 338.12,
   "kv_cache_max_tokens": 20427597,
   "kv_cache_max_concurrency": 155.8502024291498,
   "kv_cache_gpu_memory_utilization": 0.92,
   "cache_size_uncompressed_bytes": 242257920,
-  "cache_size_compressed_bytes": 17074464,
+  "cache_size_compressed_bytes": 17088681,
   "compression": "zstd -9 -T0 (multithreaded)",
-  "compress_time_seconds": 0.41,
-  "upload_time_seconds": 7.84
+  "compress_time_seconds": 0.44,
+  "upload_time_seconds": 8.38
 }