Merge branch 'update_matbench_testing' of github.com:smgjch/opencompass into update_matbench_testing

opencompass/configs/models/qwen3/lmdeploy_qwen3_0_6b.py

@@ -0,0 +1,19 @@
+from opencompass.models import TurboMindModelwithChatTemplate
+from opencompass.utils.text_postprocessors import extract_non_reasoning_content
+
+model = [
+    dict(
+        type=TurboMindModelwithChatTemplate,
+        abbr='qwen_3_0.6b_thinking-turbomind',
+        path='Qwen/Qwen3-0.6B',
+        engine_config=dict(session_len=32768, max_batch_size=16, tp=1),
+        gen_config=dict(
+            top_k=20, temperature=0.6, top_p=0.95, do_sample=True, enable_thinking=True
+        ),
+        max_seq_len=32768,
+        max_out_len=32000,
+        batch_size=16,
+        run_cfg=dict(num_gpus=1),
+        pred_postprocessor=dict(type=extract_non_reasoning_content)
+    ),
+]