Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
C++
setup test + docs: no --prefill-chunk, one plain line about the engine's own chunked prompt
Niko1221f2323d876d20
diff --git a/docs/DEEPSEEK41.md b/docs/DEEPSEEK41.mdindex 8a931c51f..32bd2a22c 100644--- a/docs/DEEPSEEK41.md+++ b/docs/DEEPSEEK41.md@@ -40,7 +40,7 @@ engram-IQ4_NL.gguf --engram-meta <UD shard 1> --dspark-gguf dspark-...-Q8_0.ggufthat packs without an error and gives garbage text), copies the shipped hot-expert profile (`data/dsv41/profile-2.00.bin` or `-2.25.bin`)into the pack as `expert-profile.bin`, checks `strata-dsv41 --pack-info` (every file present) and answers a 16-token test question(`tools/dsv41/smoke.py`, the text must name Paris). The start script passes `--dspark auto` (never `off`: `off` still loads the drafter-and its 7 GB of experts take the graphics memory); the RAM store sizes itself.+and its 7 GB of experts take the graphics memory); the RAM store sizes itself. Long prompts use the engine's fast chunked path on its own (no flag is written: the engine picks the chunk size for your cards).## Installdiff --git a/tools/test_setup_dsv41.py b/tools/test_setup_dsv41.pyindex 2e4eb7f6f..c02bb4e6e 100644--- a/tools/test_setup_dsv41.py+++ b/tools/test_setup_dsv41.py@@ -430,7 +430,7 @@ class InstallOfficial(Install):a = self.cfg["args"]self.assertEqual(a[a.index("--dspark") + 1], "auto") # never "off": that still loads the drafter and takes the VRAMself.assertNotIn("off", a)- self.assertEqual(a[a.index("--prefill-chunk") + 1], "1024")+ self.assertNotIn("--prefill-chunk", a) # the engine's own default is the layer-major prompt pathself.assertEqual(a[a.index("--dspark-conf") + 1], "0.65") # never a floor below 0.3 # the layer-major prompt pathself.assertNotIn("--ssd-cache", a) # the store sizes itselfself.assertEqual(a[a.index("--max-context") + 1], "8192")