Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

config.json

@@ -1,39 +1,70 @@
 {
-  "_attn_implementation_autoset": true,
-  "_name_or_path": "/monster/data/model/Qwen2.5-0.5B-Instruct",
   "architectures": [
     "Qwen2ForCausalLM"
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 151643,
+  "dtype": "bfloat16",
   "eos_token_id": 151645,
   "hidden_act": "silu",
   "hidden_size": 896,
   "initializer_range": 0.02,
   "intermediate_size": 4864,
+  "layer_types": [
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention",
+    "full_attention"
+  ],
   "max_position_embeddings": 32768,
   "max_window_layers": 21,
   "model_type": "qwen2",
   "num_attention_heads": 14,
   "num_hidden_layers": 24,
   "num_key_value_heads": 2,
+  "pad_token_id": 151662,
   "quantization_config": {
     "bits": 4,
     "checkpoint_format": "gptq",
-    "desc_act": true,
-    "dynamic": null,
+    "desc_act": false,
     "group_size": 128,
     "lm_head": false,
     "meta": {
-      "damp_auto_increment": 0.0025,
-      "damp_percent": 0.1,
+      "act_group_aware": true,
+      "damp_auto_increment": 0.01,
+      "damp_percent": 0.05,
+      "mse": 0.0,
       "quantizer": [
-        "gptqmodel:1.7.0"
+        "gptqmodel:5.0.0-dev0"
       ],
       "static_groups": false,
       "true_sequential": true,
-      "uri": "https://github.com/modelcloud/gptqmodel"
+      "uri": "https://github.com/modelcloud/gptqmodel",
+      "v2": false,
+      "v2_alpha": 0.25
     },
+    "pack_dtype": "int32",
+    "pack_impl": "cpu",
     "quant_method": "gptq",
     "sym": true
   },
@@ -42,8 +73,7 @@
   "rope_theta": 1000000.0,
   "sliding_window": null,
   "tie_word_embeddings": true,
-  "torch_dtype": "bfloat16",
-  "transformers_version": "4.49.0.dev0",
+  "transformers_version": "4.57.1",
   "use_cache": true,
   "use_sliding_window": false,
   "vocab_size": 151936

generation_config.json

@@ -0,0 +1,11 @@
+{
+  "bos_token_id": 151643,
+  "do_sample": true,
+  "eos_token_id": [
+    151645,
+    151643
+  ],
+  "repetition_penalty": 1.1,
+  "top_k": 20,
+  "transformers_version": "4.57.1"
+}

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:71f77eda26499bb3b6aa6a889f6852b0dd0e42fdbbe17ecee7fbfdf7ec631611
-size 731653040
+oid sha256:05d1802c2690026b6326f91a35e2d7740c44b526c602eb84d651c787ee358218
+size 459384459

quant_log.csv

@@ -1,169 +1,169 @@
-layer,module,loss,damp,time
-0,self_attn.k_proj,0.00669,0.10000,0.572
-0,self_attn.v_proj,0.00015,0.10000,0.366
-0,self_attn.q_proj,0.04556,0.10000,0.369
-0,self_attn.o_proj,0.00003,0.10000,0.365
-0,mlp.up_proj,0.57997,0.10000,0.394
-0,mlp.gate_proj,0.91966,0.10000,0.393
-0,mlp.down_proj,0.01179,0.10000,1.944
-1,self_attn.k_proj,0.06515,0.10000,0.361
-1,self_attn.v_proj,0.00657,0.10000,0.371
-1,self_attn.q_proj,0.22341,0.10000,0.374
-1,self_attn.o_proj,0.00404,0.10000,0.374
-1,mlp.up_proj,0.78053,0.10000,0.401
-1,mlp.gate_proj,1.47476,0.10000,0.398
-1,mlp.down_proj,0.01116,0.10000,1.960
-2,self_attn.k_proj,0.13747,0.10000,0.371
-2,self_attn.v_proj,0.01578,0.10000,0.367
-2,self_attn.q_proj,0.46999,0.10000,0.372
-2,self_attn.o_proj,0.00497,0.10000,0.370
-2,mlp.up_proj,1.05533,0.10000,0.400
-2,mlp.gate_proj,1.95990,0.10000,0.395
-2,mlp.down_proj,0.54380,0.10000,2.036
-3,self_attn.k_proj,0.14006,0.10000,0.365
-3,self_attn.v_proj,0.03060,0.10000,0.370
-3,self_attn.q_proj,0.59166,0.10000,0.380
-3,self_attn.o_proj,0.00675,0.10000,0.380
-3,mlp.up_proj,1.98003,0.10000,0.401
-3,mlp.gate_proj,3.62075,0.10000,0.396
-3,mlp.down_proj,0.64975,0.10000,2.032
-4,self_attn.k_proj,0.09814,0.10000,0.366
-4,self_attn.v_proj,0.04382,0.10000,0.366
-4,self_attn.q_proj,0.46673,0.10000,0.373
-4,self_attn.o_proj,0.01211,0.10000,0.377
-4,mlp.up_proj,1.58077,0.10000,0.401
-4,mlp.gate_proj,2.47939,0.10000,0.397
-4,mlp.down_proj,0.03189,0.10000,1.972
-5,self_attn.k_proj,0.12246,0.10000,0.368
-5,self_attn.v_proj,0.05167,0.10000,0.368
-5,self_attn.q_proj,0.61106,0.10000,0.373
-5,self_attn.o_proj,0.01123,0.10000,0.375
-5,mlp.up_proj,2.01079,0.10000,0.406
-5,mlp.gate_proj,4.08619,0.10000,0.396
-5,mlp.down_proj,0.03081,0.10000,1.992
-6,self_attn.k_proj,0.11447,0.10000,0.363
-6,self_attn.v_proj,0.03224,0.10000,0.361
-6,self_attn.q_proj,0.50002,0.10000,0.371
-6,self_attn.o_proj,0.01039,0.10000,0.371
-6,mlp.up_proj,1.59279,0.10000,0.403
-6,mlp.gate_proj,2.33140,0.10000,0.424
-6,mlp.down_proj,0.03129,0.10000,1.943
-7,self_attn.k_proj,0.17277,0.10000,0.363
-7,self_attn.v_proj,0.04741,0.10000,0.364
-7,self_attn.q_proj,0.70939,0.10000,0.370
-7,self_attn.o_proj,0.01649,0.10000,0.373
-7,mlp.up_proj,1.71943,0.10000,0.401
-7,mlp.gate_proj,2.22081,0.10000,0.397
-7,mlp.down_proj,0.03903,0.10000,1.962
-8,self_attn.k_proj,0.11656,0.10000,0.360
-8,self_attn.v_proj,0.04624,0.10000,0.365
-8,self_attn.q_proj,0.79887,0.10000,0.371
-8,self_attn.o_proj,0.02273,0.10000,0.368
-8,mlp.up_proj,1.43981,0.10000,0.413
-8,mlp.gate_proj,1.95721,0.10000,0.397
-8,mlp.down_proj,0.02604,0.10000,1.942
-9,self_attn.k_proj,0.33948,0.10000,0.362
-9,self_attn.v_proj,0.08356,0.10000,0.362
-9,self_attn.q_proj,1.38756,0.10000,0.373
-9,self_attn.o_proj,0.01633,0.10000,0.374
-9,mlp.up_proj,1.48946,0.10000,0.400
-9,mlp.gate_proj,1.75232,0.10000,0.396
-9,mlp.down_proj,0.02843,0.10000,1.953
-10,self_attn.k_proj,0.10033,0.10000,0.365
-10,self_attn.v_proj,0.06627,0.10000,0.364
-10,self_attn.q_proj,0.50354,0.10000,0.375
-10,self_attn.o_proj,0.02601,0.10000,0.373
-10,mlp.up_proj,1.28366,0.10000,0.400
-10,mlp.gate_proj,1.68165,0.10000,0.395
-10,mlp.down_proj,0.02453,0.10000,1.970
-11,self_attn.k_proj,0.35640,0.10000,0.364
-11,self_attn.v_proj,0.08797,0.10000,0.363
-11,self_attn.q_proj,1.44697,0.10000,0.373
-11,self_attn.o_proj,0.01708,0.10000,0.374
-11,mlp.up_proj,1.46764,0.10000,0.400
-11,mlp.gate_proj,1.59477,0.10000,0.403
-11,mlp.down_proj,0.03929,0.10000,1.959
-12,self_attn.k_proj,0.09292,0.10000,0.366
-12,self_attn.v_proj,0.06008,0.10000,0.366
-12,self_attn.q_proj,0.50941,0.10000,0.376
-12,self_attn.o_proj,0.02493,0.10000,0.407
-12,mlp.up_proj,1.36413,0.10000,0.420
-12,mlp.gate_proj,1.50805,0.10000,0.423
-12,mlp.down_proj,0.02825,0.10000,2.090
-13,self_attn.k_proj,0.17860,0.10000,0.391
-13,self_attn.v_proj,0.06740,0.10000,0.392
-13,self_attn.q_proj,0.91921,0.10000,0.375
-13,self_attn.o_proj,0.02215,0.10000,0.405
-13,mlp.up_proj,1.60880,0.10000,0.428
-13,mlp.gate_proj,1.61348,0.10000,0.419
-13,mlp.down_proj,0.03996,0.10000,1.946
-14,self_attn.k_proj,0.12914,0.10000,0.363
-14,self_attn.v_proj,0.10303,0.10000,0.360
-14,self_attn.q_proj,0.79523,0.10000,0.370
-14,self_attn.o_proj,0.04685,0.10000,0.372
-14,mlp.up_proj,1.58363,0.10000,0.395
-14,mlp.gate_proj,1.77731,0.10000,0.392
-14,mlp.down_proj,0.04859,0.10000,1.937
-15,self_attn.k_proj,0.13345,0.10000,0.362
-15,self_attn.v_proj,0.08482,0.10000,0.361
-15,self_attn.q_proj,0.72791,0.10000,0.372
-15,self_attn.o_proj,0.03135,0.10000,0.371
-15,mlp.up_proj,2.07139,0.10000,0.398
-15,mlp.gate_proj,2.14117,0.10000,0.393
-15,mlp.down_proj,0.06882,0.10000,2.049
-16,self_attn.k_proj,0.25867,0.10000,0.366
-16,self_attn.v_proj,0.23036,0.10000,0.368
-16,self_attn.q_proj,1.52680,0.10000,0.372
-16,self_attn.o_proj,0.03056,0.10000,0.404
-16,mlp.up_proj,2.59174,0.10000,0.404
-16,mlp.gate_proj,3.33505,0.10000,0.403
-16,mlp.down_proj,0.12967,0.10000,2.131
-17,self_attn.k_proj,0.15921,0.10000,0.364
-17,self_attn.v_proj,0.14702,0.10000,0.363
-17,self_attn.q_proj,1.03912,0.10000,0.379
-17,self_attn.o_proj,0.02504,0.10000,0.373
-17,mlp.up_proj,2.94703,0.10000,0.400
-17,mlp.gate_proj,4.60301,0.10000,0.399
-17,mlp.down_proj,0.10250,0.10000,2.064
-18,self_attn.k_proj,0.17395,0.10000,0.362
-18,self_attn.v_proj,0.13174,0.10000,0.361
-18,self_attn.q_proj,1.01365,0.10000,0.377
-18,self_attn.o_proj,0.02879,0.10000,0.382
-18,mlp.up_proj,2.89602,0.10000,0.416
-18,mlp.gate_proj,3.61816,0.10000,0.394
-18,mlp.down_proj,0.11206,0.10000,2.045
-19,self_attn.k_proj,0.15185,0.10000,0.365
-19,self_attn.v_proj,0.12662,0.10000,0.366
-19,self_attn.q_proj,0.91668,0.10000,0.372
-19,self_attn.o_proj,0.04191,0.10000,0.377
-19,mlp.up_proj,4.22402,0.10000,0.402
-19,mlp.gate_proj,5.07732,0.10000,0.397
-19,mlp.down_proj,0.21600,0.10000,1.942
-20,self_attn.k_proj,0.21450,0.10000,0.368
-20,self_attn.v_proj,0.46156,0.10000,0.375
-20,self_attn.q_proj,1.34480,0.10000,0.383
-20,self_attn.o_proj,0.07874,0.10000,0.373
-20,mlp.up_proj,4.72404,0.10000,0.405
-20,mlp.gate_proj,5.23060,0.10000,0.420
-20,mlp.down_proj,0.30142,0.10000,1.931
-21,self_attn.k_proj,0.25363,0.10000,0.363
-21,self_attn.v_proj,0.68723,0.10000,0.361
-21,self_attn.q_proj,1.54567,0.10000,0.371
-21,self_attn.o_proj,0.19063,0.10000,0.373
-21,mlp.up_proj,4.74235,0.10000,0.399
-21,mlp.gate_proj,5.02750,0.10000,0.393
-21,mlp.down_proj,1.48262,0.10000,1.988
-22,self_attn.k_proj,0.24376,0.10000,0.361
-22,self_attn.v_proj,0.69571,0.10000,0.368
-22,self_attn.q_proj,1.45133,0.10000,0.380
-22,self_attn.o_proj,0.05970,0.10000,0.381
-22,mlp.up_proj,4.15480,0.10000,0.406
-22,mlp.gate_proj,3.93745,0.10000,0.398
-22,mlp.down_proj,0.47177,0.10000,1.975
-23,self_attn.k_proj,0.30426,0.10000,0.368
-23,self_attn.v_proj,0.64836,0.10000,0.365
-23,self_attn.q_proj,1.41738,0.10000,0.374
-23,self_attn.o_proj,0.18444,0.10000,0.374
-23,mlp.up_proj,4.43292,0.10000,0.401
-23,mlp.gate_proj,4.59121,0.10000,0.397
-23,mlp.down_proj,0.96985,0.10000,1.943
+layer,module,loss,samples,damp,time
+0,self_attn.k_proj,0.0000001846,0.05000,0.421
+0,self_attn.v_proj,0.0000000044,0.05000,0.623
+0,self_attn.q_proj,0.0000012105,0.05000,0.651
+0,self_attn.o_proj,0.0000000008,0.05000,0.272
+0,mlp.up_proj,0.0000175088,0.05000,0.287
+0,mlp.gate_proj,0.0000274922,0.05000,0.304
+0,mlp.down_proj,0.0000003773,0.05000,1.938
+1,self_attn.k_proj,0.0000020258,0.05000,0.378
+1,self_attn.v_proj,0.0000002284,0.05000,0.662
+1,self_attn.q_proj,0.0000073043,0.05000,0.671
+1,self_attn.o_proj,0.0000001099,0.05000,0.616
+1,mlp.up_proj,0.0000257946,0.05000,0.674
+1,mlp.gate_proj,0.0000495168,0.05000,0.693
+1,mlp.down_proj,0.0000003927,0.05000,1.795
+2,self_attn.k_proj,0.0000044331,0.05000,0.442
+2,self_attn.v_proj,0.0000006006,0.05000,0.808
+2,self_attn.q_proj,0.0000161915,0.05000,0.832
+2,self_attn.o_proj,0.0000001450,0.05000,0.437
+2,mlp.gate_proj,0.0000690778,0.05000,0.259
+2,mlp.up_proj,0.0000374052,0.05000,0.706
+2,mlp.down_proj,0.0000096427,0.05000,1.414
+3,self_attn.k_proj,0.0000051747,0.05000,0.492
+3,self_attn.v_proj,0.0000012535,0.05000,0.897
+3,self_attn.q_proj,0.0000214496,0.05000,0.907
+3,self_attn.o_proj,0.0000002033,0.05000,0.422
+3,mlp.gate_proj,0.0001435908,0.05000,0.493
+3,mlp.up_proj,0.0000814998,0.05000,0.537
+3,mlp.down_proj,0.0000165190,0.05000,1.751
+4,self_attn.v_proj,0.0000015579,0.05000,0.470
+4,self_attn.q_proj,0.0000169404,0.05000,0.474
+4,self_attn.k_proj,0.0000033608,0.05000,0.710
+4,self_attn.o_proj,0.0000004269,0.05000,0.545
+4,mlp.up_proj,0.0000605056,0.05000,0.257
+4,mlp.gate_proj,0.0000924927,0.05000,0.268
+4,mlp.down_proj,0.0000012452,0.05000,2.330
+5,self_attn.k_proj,0.0000042852,0.05000,0.248
+5,self_attn.v_proj,0.0000020662,0.05000,0.388
+5,self_attn.q_proj,0.0000212661,0.05000,0.393
+5,self_attn.o_proj,0.0000003989,0.05000,0.239
+5,mlp.gate_proj,0.0001488598,0.05000,0.466
+5,mlp.up_proj,0.0000753015,0.05000,0.479
+5,mlp.down_proj,0.0000011859,0.05000,2.195
+6,self_attn.k_proj,0.0000039774,0.05000,0.668
+6,self_attn.q_proj,0.0000173516,0.05000,1.329
+6,self_attn.v_proj,0.0000011332,0.05000,1.336
+6,self_attn.o_proj,0.0000002758,0.05000,0.556
+6,mlp.up_proj,0.0000618804,0.05000,0.244
+6,mlp.gate_proj,0.0000890550,0.05000,0.289
+6,mlp.down_proj,0.0000013619,0.05000,3.632
+7,self_attn.k_proj,0.0000060954,0.05000,0.255
+7,self_attn.q_proj,0.0000250038,0.05000,1.341
+7,self_attn.v_proj,0.0000019706,0.05000,1.345
+7,self_attn.o_proj,0.0000006489,0.05000,0.269
+7,mlp.up_proj,0.0000707421,0.05000,0.258
+7,mlp.gate_proj,0.0000907406,0.05000,0.683
+7,mlp.down_proj,0.0000017962,0.05000,3.559
+8,self_attn.k_proj,0.0000045065,0.05000,0.372
+8,self_attn.v_proj,0.0000016776,0.05000,0.670
+8,self_attn.q_proj,0.0000237909,0.05000,0.693
+8,self_attn.o_proj,0.0000007376,0.05000,0.356
+8,mlp.up_proj,0.0000616151,0.05000,0.469
+8,mlp.gate_proj,0.0000799935,0.05000,0.557
+8,mlp.down_proj,0.0000011966,0.05000,3.384
+9,self_attn.k_proj,0.0000126081,0.05000,0.343
+9,self_attn.q_proj,0.0000496791,0.05000,0.533
+9,self_attn.v_proj,0.0000033377,0.05000,0.570
+9,self_attn.o_proj,0.0000005403,0.05000,0.411
+9,mlp.gate_proj,0.0000756505,0.05000,0.502
+9,mlp.up_proj,0.0000649747,0.05000,0.518
+9,mlp.down_proj,0.0000014291,0.05000,1.314
+10,self_attn.k_proj,0.0000040391,0.05000,0.408
+10,self_attn.q_proj,0.0000211797,0.05000,0.704
+10,self_attn.v_proj,0.0000029188,0.05000,0.713
+10,self_attn.o_proj,0.0000008407,0.05000,0.387
+10,mlp.gate_proj,0.0000737496,0.05000,0.456
+10,mlp.up_proj,0.0000574715,0.05000,0.465
+10,mlp.down_proj,0.0000011740,0.05000,3.047
+11,self_attn.q_proj,0.0000515761,0.05000,0.552
+11,self_attn.v_proj,0.0000037881,0.05000,0.567
+11,self_attn.k_proj,0.0000141380,0.05000,0.632
+11,self_attn.o_proj,0.0000005721,0.05000,0.704
+11,mlp.gate_proj,0.0000717162,0.05000,0.263
+11,mlp.up_proj,0.0000662663,0.05000,0.535
+11,mlp.down_proj,0.0000017422,0.05000,2.340
+12,self_attn.q_proj,0.0000219778,0.05000,0.404
+12,self_attn.v_proj,0.0000025610,0.05000,0.417
+12,self_attn.k_proj,0.0000039857,0.05000,0.663
+12,self_attn.o_proj,0.0000009195,0.05000,0.678
+12,mlp.gate_proj,0.0000669683,0.05000,0.263
+12,mlp.up_proj,0.0000601216,0.05000,0.676
+12,mlp.down_proj,0.0000013678,0.05000,2.078
+13,self_attn.q_proj,0.0000371196,0.05000,0.511
+13,self_attn.v_proj,0.0000030710,0.05000,0.517
+13,self_attn.k_proj,0.0000074217,0.05000,0.612
+13,self_attn.o_proj,0.0000007384,0.05000,0.669
+13,mlp.gate_proj,0.0000725489,0.05000,0.232
+13,mlp.up_proj,0.0000715233,0.05000,0.550
+13,mlp.down_proj,0.0000019356,0.05000,3.418
+14,self_attn.k_proj,0.0000051945,0.05000,0.446
+14,self_attn.q_proj,0.0000330241,0.05000,0.719
+14,self_attn.v_proj,0.0000047070,0.05000,0.733
+14,self_attn.o_proj,0.0000018053,0.05000,0.223
+14,mlp.up_proj,0.0000730560,0.05000,0.249
+14,mlp.gate_proj,0.0000833288,0.05000,0.251
+14,mlp.down_proj,0.0000024172,0.05000,1.230
+15,self_attn.k_proj,0.0000056382,0.05000,0.430
+15,self_attn.v_proj,0.0000039362,0.05000,0.856
+15,self_attn.q_proj,0.0000289588,0.05000,0.860
+15,self_attn.o_proj,0.0000010586,0.05000,0.353
+15,mlp.gate_proj,0.0000967614,0.05000,0.617
+15,mlp.up_proj,0.0000926957,0.05000,0.702
+15,mlp.down_proj,0.0000031419,0.05000,2.435
+16,self_attn.k_proj,0.0000109810,0.05000,0.443
+16,self_attn.v_proj,0.0000096741,0.05000,0.734
+16,self_attn.q_proj,0.0000628785,0.05000,0.748
+16,self_attn.o_proj,0.0000011271,0.05000,0.389
+16,mlp.gate_proj,0.0001423836,0.05000,0.274
+16,mlp.up_proj,0.0001109843,0.05000,0.730
+16,mlp.down_proj,0.0000060371,0.05000,1.237
+17,self_attn.k_proj,0.0000060039,0.05000,0.309
+17,self_attn.q_proj,0.0000400760,0.05000,0.681
+17,self_attn.v_proj,0.0000059069,0.05000,0.682
+17,self_attn.o_proj,0.0000008255,0.05000,0.386
+17,mlp.up_proj,0.0001233144,0.05000,0.384
+17,mlp.gate_proj,0.0001934214,0.05000,0.412
+17,mlp.down_proj,0.0000050482,0.05000,2.300
+18,self_attn.q_proj,0.0000411873,0.05000,0.357
+18,self_attn.v_proj,0.0000059423,0.05000,0.369
+18,self_attn.k_proj,0.0000077203,0.05000,0.678
+18,self_attn.o_proj,0.0000010869,0.05000,0.599
+18,mlp.up_proj,0.0001260548,0.05000,0.258
+18,mlp.gate_proj,0.0001582911,0.05000,0.264
+18,mlp.down_proj,0.0000059834,0.05000,2.446
+19,self_attn.k_proj,0.0000063490,0.05000,0.272
+19,self_attn.q_proj,0.0000380589,0.05000,0.666
+19,self_attn.v_proj,0.0000057355,0.05000,0.670
+19,self_attn.o_proj,0.0000013602,0.05000,0.247
+19,mlp.up_proj,0.0001832222,0.05000,0.367
+19,mlp.gate_proj,0.0002190037,0.05000,0.372
+19,mlp.down_proj,0.0000113082,0.05000,2.235
+20,self_attn.k_proj,0.0000078660,0.05000,0.290
+20,self_attn.q_proj,0.0000533177,0.05000,0.443
+20,self_attn.v_proj,0.0000177706,0.05000,0.459
+20,self_attn.o_proj,0.0000024812,0.05000,0.569
+20,mlp.up_proj,0.0002066750,0.05000,0.240
+20,mlp.gate_proj,0.0002273082,0.05000,0.717
+20,mlp.down_proj,0.0000155725,0.05000,2.551
+21,self_attn.k_proj,0.0000082325,0.05000,0.232
+21,self_attn.q_proj,0.0000596139,0.05000,1.457
+21,self_attn.v_proj,0.0000242640,0.05000,1.457
+21,self_attn.o_proj,0.0000054829,0.05000,0.257
+21,mlp.up_proj,0.0002139384,0.05000,0.277
+21,mlp.gate_proj,0.0002226604,0.05000,0.695
+21,mlp.down_proj,0.0000279960,0.05000,2.661
+22,self_attn.k_proj,0.0000075570,0.05000,0.458
+22,self_attn.v_proj,0.0000210302,0.05000,0.811
+22,self_attn.q_proj,0.0000556213,0.05000,0.815
+22,self_attn.o_proj,0.0000020455,0.05000,0.256
+22,mlp.up_proj,0.0001812818,0.05000,0.505
+22,mlp.gate_proj,0.0001728906,0.05000,0.536
+22,mlp.down_proj,0.0000204518,0.05000,1.919
+23,self_attn.k_proj,0.0000085684,0.05000,0.263
+23,self_attn.v_proj,0.0000168603,0.05000,0.415
+23,self_attn.q_proj,0.0000655303,0.05000,0.421
+23,self_attn.o_proj,0.0000053465,0.05000,0.242
+23,mlp.gate_proj,0.0001968066,0.05000,0.261
+23,mlp.up_proj,0.0001895685,0.05000,0.267
+23,mlp.down_proj,0.0000390435,0.05000,1.745

quantize_config.json

@@ -1,21 +1,25 @@
 {
   "bits": 4,
-  "dynamic": null,
   "group_size": 128,
-  "desc_act": true,
+  "desc_act": false,
   "sym": true,
   "lm_head": false,
   "quant_method": "gptq",
   "checkpoint_format": "gptq",
+  "pack_dtype": "int32",
   "meta": {
     "quantizer": [
-      "gptqmodel:1.7.0"
+      "gptqmodel:5.0.0-dev0"
     ],
     "uri": "https://github.com/modelcloud/gptqmodel",
-    "damp_percent": 0.1,
-    "damp_auto_increment": 0.0025,
+    "damp_percent": 0.05,
+    "damp_auto_increment": 0.01,
     "static_groups": false,
     "true_sequential": true,
-    "mse": 2.4
-  }
+    "mse": 0.0,
+    "v2": false,
+    "v2_alpha": 0.25,
+    "act_group_aware": true
+  },
+  "pack_impl": "cpu"
 }

special_tokens_map.json

@@ -21,5 +21,5 @@
     "rstrip": false,
     "single_word": false
   },
-  "pad_token": "<unk>"
+  "pad_token": "<|fim_pad|>"
 }

tokenizer_config.json

@@ -160,7 +160,7 @@
       "normalized": false,
       "rstrip": false,
       "single_word": false,
-      "special": false
+      "special": true
     },
     "151663": {
       "content": "<|repo_name|>",
@@ -195,13 +195,13 @@
     "<|video_pad|>"
   ],
   "bos_token": null,
-  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
   "clean_up_tokenization_spaces": false,
   "eos_token": "<|im_end|>",
   "errors": "replace",
+  "extra_special_tokens": {},
   "model_max_length": 131072,
-  "pad_token": "<|endoftext|>",
+  "pad_token": "<|fim_pad|>",
   "split_special_tokens": false,
   "tokenizer_class": "Qwen2Tokenizer",
   "unk_token": null
-}
+}