From 0732f6003babc9fb6f656458073a42b4dbe197bb Mon Sep 17 00:00:00 2001 From: bong-water-water-bong Date: Tue, 10 Nov 2026 16:21:46 -0400 Subject: [PATCH] census: MicroLlamaForCausalLM is report-only -- its model_type is llama, which is already mapped (engine#1922) The daily census flags `MicroLlamaForCausalLM` UNCOVERED because that class string is in no registry row. Its own config.json says why that is not a capability gap: "architectures": ["MicroLlamaForCausalLM"], "model_type": "llama", vocab 1024, hidden 192, 6 layers, 6 heads, 256 positions model_type is `llama`, and the registry already maps it (LlamaForCausalLM -> gguf llama, hrx+npu, npu_model_type llama). The NPU path matches on model_type, not on the class name (tools/registry_build.py says so), so "UNCOVERED" here means the census could not find the CLASS STRING, not that the engine cannot serve the model. Recorded the reviewed disposition the way #1972 recorded the DeepSeek V4.1 Flash SuperExpert one: a `watch: "report"` row in the curated registry/significant.json (not the generated architectures.json, which tools/registry_build.py --check rejects a hand edit of), with `mapped_ok` naming the mapping that already exists. Not amended into architectures.json, because adding a class row for a name that already resolves would record the census search strategy rather than the engine capability. Checks: tools/registry_build.py --check-gaps -> "none of the 38 reviewed significant classes is mapped without a reason"; tests/census_watch_test.py -> all checks passed; copyright rc=0. --- registry/significant.json | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/registry/significant.json b/registry/significant.json index 75ac3b192..12cbc93ef 100644 --- a/registry/significant.json +++ b/registry/significant.json @@ -172,6 +172,12 @@ "model_type": "deepseek_v4_flash_superexpert", "family": "DeepSeek V4.1 Flash SuperExpert, published as remote code and with AW-SECD-compressed weights (8.5 GB, claimed 0.962 retention); MLA-style geometry (kv_lora_rank 512, q_lora_rank 1536, qk_rope_head_dim 64, v_head_dim 128, 32 heads / 16 KV, IM 4096). Not an alias of deepseek_v41: nothing at the pin decodes AW-SECD weights, so mapping it would name a backend that cannot read it", "watch": "report" + }, + "MicroLlamaForCausalLM": { + "model_type": "llama", + "family": "MicroLlama (Climb-Ai/Climb-1-Toki: vocab 1024, hidden 192, 6 heads, 6 layers, 256 positions)", + "watch": "report", + "mapped_ok": "engine#1922: model_type is `llama`, which the registry already maps (LlamaForCausalLM -> gguf llama, hrx+npu). Only the architectures[0] CLASS NAME is new, and the NPU path matches on model_type, not on the class name (tools/registry_build.py). Reported, not alerted: a renamed fine-tune of a supported family is not a coverage gap." } } }