compiled artefacts from mac_arm64 (cpu)

README.md

@@ -279,8 +279,9 @@ Compiled engines, named per **runtime × OS × GPU arch**, plus the ONNX for sta
   `classifier.dense` / `classifier.out_proj` randomly initialised and ignore ours. Load the
   encoder with `AutoModel`, mean-pool the last hidden state over the tokens equal to each marker
   id (`config.json` → `ultracode.entity_marker_ids`, 250002 and 250004), concatenate the two and
-  apply the classifier. Done that way it matches the ONNX to 2e-6. There is no MLX build of this
-  model: the Apple runtime implements no entity-pool head.
+  apply the classifier. Done that way it matches the ONNX to 2e-6. The Apple (MLX) build is
+  prepared from this pair: the UltraCode MLX runtime implements this entity-pool head, with the
+  marker ids carried in the prepared file's header.
 
 FP16 rather than INT8: this architecture's activation outliers make per-tensor INT8 calibration
 lossy, and FP16 costs nothing on any GPU that can run it.

manifest.json

@@ -12,6 +12,6 @@
     "tvm": "tvm0.25",
     "mlx_dir": "model_gpu_mlx0.32"
   },
-  "compiled_at": "2026-09-17T13:27:03Z",
+  "compiled_at": "2026-09-17T18:48:45Z",
   "compiled_by": "models/_compile"
 }

model_gpu_mlx0.32/config.json

@@ -0,0 +1,57 @@
+{
+  "add_cross_attention": false,
+  "architectures": [
+    "XLMRobertaForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "custom_head": "entity_start_marker_meanmask (concat E1,E2 -> Linear); [CLS] is not read",
+  "dtype": "float32",
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 384,
+  "id2label": {
+    "0": "NO_RELATION",
+    "1": "semantically_similar_to",
+    "2": "supersedes_or_conflicts",
+    "3": "depends_on"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 1536,
+  "is_decoder": false,
+  "label2id": {
+    "NO_RELATION": 0,
+    "semantically_similar_to": 1,
+    "supersedes_or_conflicts": 2,
+    "depends_on": 3
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "xlm-roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "num_labels": 4,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "sbert_ce_default_activation_function": "torch.nn.modules.linear.Identity",
+  "tie_word_embeddings": true,
+  "transformers_version": "5.1.0",
+  "type_vocab_size": 1,
+  "ultracode_note": "Inference artefact is model.onnx (inputs: input_ids, attention_mask; output: logits[batch, 4]). Feed the pair as (empty query, marked passage).",
+  "use_cache": true,
+  "vocab_size": 250006,
+  "ultracode": {
+    "format": "raw-safetensors/1",
+    "exported_from": "model.onnx",
+    "onnx_sha256": "73556d636633563bd0688d52899c475ff61a48f0687e8c48703ee4b9bc74d135",
+    "head": "entity_pool",
+    "position_ids": "roberta",
+    "pooling": "mean of the last hidden state over the tokens equal to each marker id; concat(E1, E2) -> classifier",
+    "entity_marker_ids": [
+      250002,
+      250004
+    ]
+  }
+}

model_gpu_mlx0.32/mlx_config.json

@@ -0,0 +1,7 @@
+{
+  "quantize": "none",
+  "format": "mlx-safetensors",
+  "prep_version": 2,
+  "head": "entity_pool",
+  "num_attention_heads": 12
+}

model_gpu_mlx0.32/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc022e30f7821a44c9f0bc5e24dc5b3e4a5760a0dd3019f59f6d8ee398a453f7
+size 470014008