Sync model repo (text/metadata)

README.md

@@ -68,7 +68,7 @@ Performance was measured on the reference configuration below. Results are inten
 | Runtime | ONNX Runtime 1.27.0 |
 | Backend / delegate | MLAS, KleidiAI |
 | Batch size | 1 |
-| Precision | INT4 groupwise asymmetric weights (GPTQ, group size 32), INT8 per-token dynamic activations, INT8 per-row embeddings, INT8 LM head (scheme W4A8_dyn_emb_int8_lmhead_int8) |
+| Precision | INT4 groupwise asymmetric weights (GPTQ, group size 32), INT8 per-token dynamic activations, INT8 per-row embeddings, INT8 LM head (scheme W4A8_dyn_emb_int8) |
 | Runs | 20 warmup runs + 100 measured runs |
 
 **Performance results**
@@ -94,7 +94,7 @@ Accuracy was evaluated using the same preprocessing, input resolution, and evalu
 | Field | Value |
 |---|---|
 | Dataset | [HellaSwag](https://huggingface.co/datasets/Rowan/hellaswag) |
-| Split | 0-shot |
+| Split | validation (0-shot) |
 | Number of samples | 8000 |
 | Metric(s) | HellaSwag accuracy, acc_norm convention (character-length-normalized) |
 | Evaluation runtime | ONNX Runtime 1.27.0 |
@@ -116,7 +116,7 @@ For this release, Arm used:
 | Optimization area | Applied? | Notes |
 |---|---|---|
 | Model conversion | Yes | Converted to ONNX in the ONNX Runtime GenAI graph layout, with attention fused into a single GroupQueryAttention op and the KV cache wired internally |
-| Quantization | Yes | GPTQ weight-only quantization to INT4, asymmetric per-group with group size 32, calibrated on 256 randomly selected WikiText2 samples; INT8 per-token dynamic activations; INT8 per-row embedding table; INT8 LM head protected by k_quant_last (scheme W4A8_dyn_emb_int8_lmhead_int8) |
+| Quantization | Yes | GPTQ weight-only quantization to INT4, asymmetric per-group with group size 32, calibrated on 256 randomly selected WikiText2 samples; INT8 per-token dynamic activations; INT8 per-row embedding table; INT8 LM head protected by k_quant_last (scheme W4A8_dyn_emb_int8) |
 | Runtime/backend selection | Yes | ONNX Runtime 1.27.0 CPU execution provider, with MLAS and KleidiAI kernels |
 | Accuracy validation | Yes | Compared against the original model or published baseline |
 | Performance validation | Yes | Measured on the reference Arm platform |