Clarify INT8 quantization in card and metadata

README.md

@@ -1,7 +1,7 @@
 ---
 license: apache-2.0
-base_model:
-- Nanbeige/Nanbeige4.2-3B
+base_model: Nanbeige/Nanbeige4.2-3B
+base_model_relation: quantized
 pipeline_tag: text-generation
 library_name: coreai
 tags:
@@ -9,17 +9,35 @@ tags:
 - apple-silicon
 - recurrent-transformer
 - text-generation
+- quantized
+- int8
 ---
 
-# Nanbeige4.2-3B — Core AI
+# Nanbeige4.2-3B — Core AI INT8
 
-Core AI conversion of [`Nanbeige/Nanbeige4.2-3B`](https://huggingface.co/Nanbeige/Nanbeige4.2-3B), pinned to source revision [`5ff54fb7ed86ce8e216d78bff5417ab9981de3d4`](https://huggingface.co/Nanbeige/Nanbeige4.2-3B/tree/5ff54fb7ed86ce8e216d78bff5417ab9981de3d4).
+**INT8 weight-quantized Core AI LanguageBundle** converted from [`Nanbeige/Nanbeige4.2-3B`](https://huggingface.co/Nanbeige/Nanbeige4.2-3B), pinned to source revision [`5ff54fb7ed86ce8e216d78bff5417ab9981de3d4`](https://huggingface.co/Nanbeige/Nanbeige4.2-3B/tree/5ff54fb7ed86ce8e216d78bff5417ab9981de3d4).
 
-Ported by [Vadim Smirnov (@ukint-vs)](https://github.com/ukint-vs). Source, reproducible conversion recipe, parity gates, and the full optimization report are tracked in [john-rocky/coreai-model-zoo#5](https://github.com/john-rocky/coreai-model-zoo/issues/5).
+This is a Core AI `.aimodel` bundle for Apple silicon. It is not a Transformers, GGUF, or MLX checkpoint.
+
+Ported by [Vadim Smirnov (@ukint-vs)](https://github.com/ukint-vs). Source, reproducible conversion recipe, parity gates, and the optimization report are tracked in [john-rocky/coreai-model-zoo#5](https://github.com/john-rocky/coreai-model-zoo/issues/5).
+
+## Format
+
+| Property | Value |
+|---|---|
+| Published variant | **INT8 only** |
+| Body weights | int8, per-block 32 (`int8hu`) |
+| Untied LM head | symmetric absmax int8, per-block 32 (`--head-sym`) |
+| Activations / KV cache | fp16 |
+| Decode input | static sequence length 1 (`--static-ids`) |
+| Bundle size | approximately 4.59 GiB |
+| Verified context | 4,096 tokens |
+
+The int4 and mixed int4/int8 candidates failed the reasoning quality gate and are intentionally not published.
 
 ## Architecture
 
-Nanbeige4.2 is a looped Llama: **22 physical transformer blocks execute twice**, with RMS normalization after each pass and independent KV history. The bundle therefore has **44 executed/cache layers while storing one copy of the 22 physical blocks**.
+Nanbeige4.2 is a looped Llama: **22 physical transformer blocks execute twice**, with RMS normalization after each pass and independent KV history. The bundle therefore has **44 executed/cache layers while storing and quantizing one copy of the 22 physical blocks**.
 
 ## Bundle
 
@@ -33,7 +51,7 @@ gpu-pipelined/
     tokenizer/
 ```
 
-Configuration: static-S=1, int8 body per-block-32, symmetric absmax int8 untied head, 4,096-token verified context. The embedded tokenizer includes the vendor `chat_template.jinja` with thinking enabled and disabled modes. The `.aimodel` producer is `coreai-core 1.0.0b2`.
+The embedded tokenizer includes the vendor `chat_template.jinja` with thinking enabled and disabled modes. The `.aimodel` producer is `coreai-core 1.0.0b2`.
 
 ## Verification
 
@@ -41,7 +59,6 @@ Configuration: static-S=1, int8 body per-block-32, symmetric absmax int8 untied
 - Quantized Core AI engine: token-exact vs fp32 on 24-token Paris, 16-token freezing-point, and 64-token decimal-reasoning gates.
 - M4 Max Release benchmark, prompt 128 / generation 256 / 3 runs: **47.37 prefill / 46.35 decode tok/s**.
 - 4,096-token boundary: **29.83 prefill / 32.80 decode tok/s**, 9.17 GiB peak RSS, zero swaps.
-- int4 and mixed int4/int8 candidates failed the reasoning quality gate and are intentionally not published.
 - iPhone `h18p` device acceptance is pending; no iPhone performance claim is made.
 
 ## License

gpu-pipelined/nanbeige4_2_3b_decode_int8hu_block32_sym_s1/metadata.json

@@ -21,9 +21,11 @@
     "hf_model_id": "Nanbeige/Nanbeige4.2-3B",
     "hf_revision": "5ff54fb7ed86ce8e216d78bff5417ab9981de3d4"
   },
-  "compression": null,
+  "compression": {
+    "scheme": "int8hu"
+  },
   "compilation": {
     "date": "2026-07-22T15:06:28.708628+00:00",
     "targets": []
   }
-}
+}