raw float weights: model.safetensors + config.json, exported from the shipped model.onnx

README.md

@@ -254,6 +254,11 @@ Flat layout, named per **runtime × GPU arch × OS** (single-profile — no leng
 - **ONNX** — `model.onnx` FP32 dynamic-shape (the build source, and standalone `onnxruntime` /
   `optimum` use), `model_static.onnx` (seq pinned to 128, the engine-build input), and
   `model_int8qdt.onnx` (INT8 Q/DQ).
+- **Raw weights** — `model.safetensors` + `config.json`, the FP32 weights in the
+  `XLMRobertaForTokenClassification` layout with the 17 labels in the config:
+  `AutoModelForTokenClassification` loads them with no missing or unexpected keys, and its
+  per-token logits match the ONNX to 1e-5. Calibrate them with `temperature.json` exactly as
+  in the snippet above.
 
 No MLX build: this model has no Apple-GPU consumer today.
 

config.json

@@ -0,0 +1,69 @@
+{
+  "architectures": [
+    "XLMRobertaForTokenClassification"
+  ],
+  "model_type": "xlm-roberta",
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "initializer_range": 0.02,
+  "num_attention_heads": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "vocab_size": 250002,
+  "hidden_size": 384,
+  "num_hidden_layers": 12,
+  "intermediate_size": 1536,
+  "max_position_embeddings": 514,
+  "type_vocab_size": 1,
+  "layer_norm_eps": 1e-05,
+  "id2label": {
+    "0": "O",
+    "1": "B-component",
+    "2": "I-component",
+    "3": "B-api_endpoint",
+    "4": "I-api_endpoint",
+    "5": "B-algorithm",
+    "6": "I-algorithm",
+    "7": "B-data_structure",
+    "8": "I-data_structure",
+    "9": "B-config",
+    "10": "I-config",
+    "11": "B-tool",
+    "12": "I-tool",
+    "13": "B-file_path",
+    "14": "I-file_path",
+    "15": "B-metric",
+    "16": "I-metric"
+  },
+  "label2id": {
+    "O": 0,
+    "B-component": 1,
+    "I-component": 2,
+    "B-api_endpoint": 3,
+    "I-api_endpoint": 4,
+    "B-algorithm": 5,
+    "I-algorithm": 6,
+    "B-data_structure": 7,
+    "I-data_structure": 8,
+    "B-config": 9,
+    "I-config": 10,
+    "B-tool": 11,
+    "I-tool": 12,
+    "B-file_path": 13,
+    "I-file_path": 14,
+    "B-metric": 15,
+    "I-metric": 16
+  },
+  "num_labels": 17,
+  "ultracode": {
+    "format": "raw-safetensors/1",
+    "exported_from": "model.onnx",
+    "onnx_sha256": "5a317a02eaccef10ed9bf494b768edfb64e651e112af90ec0a38e974b8d05269",
+    "head": "token_cls",
+    "position_ids": "roberta",
+    "pooling": "none: per-token classifier"
+  }
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e96f32e7f39e1fa4d886807e668b02c8adc193d492b51d4e22b042af2eee22f
+size 470021692