Hugging Face's logo

ATL-Machine
/
test

Safetensors
qwen3
Model card Files
xet
 Community
ATL-Machine commited on
Commit
72ea1ea
·
verified ·
1 Parent(s): 7c6da3b

Upload model (bfloat16) from script

Browse files
Files changed (8) hide show
  1. chat_template.jinja +81 -35
  2. config.json +67 -69
  3. generation_config.json +11 -3
  4. model-00001-of-00002.safetensors +1 -1
  5. model-00002-of-00002.safetensors +1 -1
  6. model.safetensors.index.json +2 -2
  7. special_tokens_map.json +1 -1
  8. tokenizer_config.json +4 -6
chat_template.jinja CHANGED
@@ -1,42 +1,88 @@
1
-
2
- {%- set has_system = messages and messages[0].role == 'system' and messages[0].content is string %}
3
- {%- if has_system %}
4
- {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
5
- {%- endif %}
6
-
7
- {%- if messages and messages[-1].role == 'assistant' %}
8
- {# Find the user message immediately before the last assistant (scan backwards skipping last). #}
9
- {%- set ns = namespace(seen_last=false, user_before_last_asst=none) %}
10
- {%- for m in messages[::-1] %}
11
- {%- if not ns.seen_last %}
12
- {%- set ns.seen_last = true %}
13
- {%- elif ns.user_before_last_asst is none and m.role == 'user' and m.content is string %}
14
- {%- set ns.user_before_last_asst = m.content %}
15
- {%- endif %}
16
  {%- endfor %}
17
- {%- if ns.user_before_last_asst is not none %}
18
- {{- '<|im_start|>user\n' + ns.user_before_last_asst + '<|im_end|>\n' }}
 
 
19
  {%- endif %}
20
-
21
- {%- set content = '' %}
22
- {%- if messages[-1].content is string %}
23
- {%- set content = messages[-1].content %}
 
 
 
24
  {%- endif %}
25
- {%- set reasoning_content = '' %}
26
- {%- if messages[-1].reasoning_content is string %}
27
- {%- set reasoning_content = messages[-1].reasoning_content %}
 
28
  {%- else %}
29
- {%- if '</think>' in content %}
30
- {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
31
- {%- set content = content.split('</think>')[-1].lstrip('\n') %}
32
- {%- endif %}
33
  {%- endif %}
34
- {%- if reasoning_content %}
35
- {{- '<|im_start|>assistant\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') + '<|im_end|>\n' }}
36
- {%- else %}
37
- {{- '<|im_start|>assistant\n' + content + '<|im_end|>\n' }}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38
  {%- endif %}
39
-
40
- {%- elif messages and messages[-1].role == 'user' and messages[-1].content is string %}
41
- {{- '<|im_start|>user\n' + messages[-1].content + '<|im_end|>\n<|im_start|>assistant\n' }}
42
  {%- endif %}
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0].role == 'system' %}
4
+ {{- messages[0].content + '\n\n' }}
5
+ {%- endif %}
6
+ {{- "# 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>" }}
7
+ {%- for tool in tools %}
8
+ {{- "\n" }}
9
+ {{- tool | tojson }}
 
 
 
 
 
 
10
  {%- endfor %}
11
+ {{- "\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" }}
12
+ {%- else %}
13
+ {%- if messages[0].role == 'system' %}
14
+ {{- '<|im_start|>system\n' + messages[0].content + '<|im_end|>\n' }}
15
  {%- endif %}
16
+ {%- endif %}
17
+ {%- set ns = namespace(multi_step_tool=true, last_query_index=messages|length - 1) %}
18
+ {%- for message in messages[::-1] %}
19
+ {%- set index = (messages|length - 1) - loop.index0 %}
20
+ {%- if ns.multi_step_tool and message.role == "user" and message.content is string and not(message.content.startswith('<tool_response>') and message.content.endswith('</tool_response>')) %}
21
+ {%- set ns.multi_step_tool = false %}
22
+ {%- set ns.last_query_index = index %}
23
  {%- endif %}
24
+ {%- endfor %}
25
+ {%- for message in messages %}
26
+ {%- if message.content is string %}
27
+ {%- set content = message.content %}
28
  {%- else %}
29
+ {%- set content = '' %}
 
 
 
30
  {%- endif %}
31
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) %}
32
+ {{- '<|im_start|>' + message.role + '\n' + content + '<|im_end|>' + '\n' }}
33
+ {%- elif message.role == "assistant" %}
34
+ {%- set reasoning_content = '' %}
35
+ {%- if message.reasoning_content is string %}
36
+ {%- set reasoning_content = message.reasoning_content %}
37
+ {%- else %}
38
+ {%- if '</think>' in content %}
39
+ {%- set reasoning_content = content.split('</think>')[0].rstrip('\n').split('<think>')[-1].lstrip('\n') %}
40
+ {%- set content = content.split('</think>')[-1].lstrip('\n') %}
41
+ {%- endif %}
42
+ {%- endif %}
43
+ {%- generation -%}
44
+ {%- if loop.index0 > ns.last_query_index %}
45
+ {%- if loop.last or (not loop.last and reasoning_content) %}
46
+ {{- '<|im_start|>' + message.role + '\n<think>\n' + reasoning_content.strip('\n') + '\n</think>\n\n' + content.lstrip('\n') }}
47
+ {%- else %}
48
+ {{- '<|im_start|>' + message.role + '\n' + content }}
49
+ {%- endif %}
50
+ {%- else %}
51
+ {{- '<|im_start|>' + message.role + '\n' + content }}
52
+ {%- endif %}
53
+ {%- if message.tool_calls %}
54
+ {%- for tool_call in message.tool_calls %}
55
+ {%- if (loop.first and content) or (not loop.first) %}
56
+ {{- '\n' }}
57
+ {%- endif %}
58
+ {%- if tool_call.function %}
59
+ {%- set tool_call = tool_call.function %}
60
+ {%- endif %}
61
+ {{- '<tool_call>\n{"name": "' }}
62
+ {{- tool_call.name }}
63
+ {{- '", "arguments": ' }}
64
+ {%- if tool_call.arguments is string %}
65
+ {{- tool_call.arguments }}
66
+ {%- else %}
67
+ {{- tool_call.arguments | tojson }}
68
+ {%- endif %}
69
+ {{- '}\n</tool_call>' }}
70
+ {%- endfor %}
71
+ {%- endif %}
72
+ {{- '<|im_end|>\n' }}
73
+ {%- endgeneration -%}
74
+ {%- elif message.role == "tool" %}
75
+ {%- if loop.first or (messages[loop.index0 - 1].role != "tool") %}
76
+ {{- '<|im_start|>user' }}
77
+ {%- endif %}
78
+ {{- '\n<tool_response>\n' }}
79
+ {{- content }}
80
+ {{- '\n</tool_response>' }}
81
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
82
+ {{- '<|im_end|>\n' }}
83
+ {%- endif %}
84
  {%- endif %}
85
+ {%- endfor %}
86
+ {%- if add_generation_prompt %}
87
+ {{- '<|im_start|>assistant\n<think>\n' }}
88
  {%- endif %}
config.json CHANGED
@@ -1,70 +1,68 @@
1
  {
2
- "architectures": [
3
- "Qwen3ForCausalLM"
4
- ],
5
- "attention_bias": false,
6
- "attention_dropout": 0.0,
7
- "torch_dtype": "bfloat16",
8
- "eos_token_id": 151645,
9
- "head_dim": 128,
10
- "hidden_act": "silu",
11
- "hidden_size": 2560,
12
- "initializer_range": 0.02,
13
- "intermediate_size": 9728,
14
- "layer_types": [
15
- "full_attention",
16
- "full_attention",
17
- "full_attention",
18
- "full_attention",
19
- "full_attention",
20
- "full_attention",
21
- "full_attention",
22
- "full_attention",
23
- "full_attention",
24
- "full_attention",
25
- "full_attention",
26
- "full_attention",
27
- "full_attention",
28
- "full_attention",
29
- "full_attention",
30
- "full_attention",
31
- "full_attention",
32
- "full_attention",
33
- "full_attention",
34
- "full_attention",
35
- "full_attention",
36
- "full_attention",
37
- "full_attention",
38
- "full_attention",
39
- "full_attention",
40
- "full_attention",
41
- "full_attention",
42
- "full_attention",
43
- "full_attention",
44
- "full_attention",
45
- "full_attention",
46
- "full_attention",
47
- "full_attention",
48
- "full_attention",
49
- "full_attention",
50
- "full_attention"
51
- ],
52
- "max_position_embeddings": 40960,
53
- "max_window_layers": 36,
54
- "model_type": "qwen3",
55
- "num_attention_heads": 32,
56
- "num_hidden_layers": 36,
57
- "num_key_value_heads": 8,
58
- "pad_token_id": 151654,
59
- "rms_norm_eps": 1e-06,
60
- "rope_scaling": null,
61
- "rope_theta": 1000000,
62
- "sliding_window": null,
63
- "tie_word_embeddings": true,
64
- "transformers_version": "4.57.6",
65
- "unsloth_fixed": true,
66
- "unsloth_version": "2026.1.2",
67
- "use_cache": true,
68
- "use_sliding_window": false,
69
- "vocab_size": 151936
70
- }
 
1
  {
2
+ "architectures": [
3
+ "Qwen3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "dtype": "bfloat16",
8
+ "eos_token_id": 151645,
9
+ "head_dim": 128,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 2560,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 9728,
14
+ "layer_types": [
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention",
42
+ "full_attention",
43
+ "full_attention",
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention"
51
+ ],
52
+ "max_position_embeddings": 262144,
53
+ "max_window_layers": 36,
54
+ "model_type": "qwen3",
55
+ "num_attention_heads": 32,
56
+ "num_hidden_layers": 36,
57
+ "num_key_value_heads": 8,
58
+ "pad_token_id": 151643,
59
+ "rms_norm_eps": 1e-06,
60
+ "rope_scaling": null,
61
+ "rope_theta": 5000000,
62
+ "sliding_window": null,
63
+ "tie_word_embeddings": true,
64
+ "transformers_version": "4.57.1",
65
+ "use_cache": false,
66
+ "use_sliding_window": false,
67
+ "vocab_size": 151936
68
+ }
 
 
generation_config.json CHANGED
@@ -1,4 +1,12 @@
1
  {
2
- "max_length": 131072,
3
- "eos_token_id": 163586
4
- }
 
 
 
 
 
 
 
 
 
1
  {
2
+ "do_sample": true,
3
+ "eos_token_id": [
4
+ 151645,
5
+ 151643
6
+ ],
7
+ "pad_token_id": 151643,
8
+ "temperature": 0.6,
9
+ "top_k": 20,
10
+ "top_p": 0.95,
11
+ "transformers_version": "4.57.1"
12
+ }
model-00001-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:712c6834f650644193df0f465e45e8fed98cd49718912047f4eebeef64b421c4
3
  size 4967215360
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fd1d9fefc5b28e6d31423995a8d787e28502b91bba181ba187c8426987ac7937
3
  size 4967215360
model-00002-of-00002.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:806bf08e28549f927cdf9912d01c9ac5ca0f7beb1e6ca417bd3d3f939cc4da2e
3
  size 3077766632
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4a8a54258f9dc5649c116e07f96de92d367fb11d1fdb470c92dddab04b447cd
3
  size 3077766632
model.safetensors.index.json CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e36bba5af4706cfd22ecf7eedb5fd8f4d2559b92f2414b932651b9ed6495921a
3
- size 32819
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06b3d5319b6d76d1a4a2433419180016cfd54ed62d086a5e6567a809f8c82634
3
+ size 32855
special_tokens_map.json CHANGED
@@ -22,7 +22,7 @@
22
  "single_word": false
23
  },
24
  "pad_token": {
25
- "content": "<|vision_pad|>",
26
  "lstrip": false,
27
  "normalized": false,
28
  "rstrip": false,
 
22
  "single_word": false
23
  },
24
  "pad_token": {
25
+ "content": "<|endoftext|>",
26
  "lstrip": false,
27
  "normalized": false,
28
  "rstrip": false,
tokenizer_config.json CHANGED
@@ -231,11 +231,9 @@
231
  "eos_token": "<|im_end|>",
232
  "errors": "replace",
233
  "extra_special_tokens": {},
234
- "model_max_length": 40960,
235
- "pad_token": "<|vision_pad|>",
236
- "padding_side": "left",
237
  "split_special_tokens": false,
238
  "tokenizer_class": "Qwen2Tokenizer",
239
- "unk_token": null,
240
- "chat_template": "\n{%- set has_system = messages and messages[0].role == 'system' and messages[0].content is string %}\n{%- if has_system %}\n{{- '<|im_start|>system\\n' + messages[0].content + '<|im_end|>\\n' }}\n{%- endif %}\n\n{%- if messages and messages[-1].role == 'assistant' %}\n {# Find the user message immediately before the last assistant (scan backwards skipping last). #}\n {%- set ns = namespace(seen_last=false, user_before_last_asst=none) %}\n {%- for m in messages[::-1] %}\n {%- if not ns.seen_last %}\n {%- set ns.seen_last = true %}\n {%- elif ns.user_before_last_asst is none and m.role == 'user' and m.content is string %}\n {%- set ns.user_before_last_asst = m.content %}\n {%- endif %}\n {%- endfor %}\n {%- if ns.user_before_last_asst is not none %}\n{{- '<|im_start|>user\\n' + ns.user_before_last_asst + '<|im_end|>\\n' }}\n {%- endif %}\n\n {%- set content = '' %}\n {%- if messages[-1].content is string %}\n {%- set content = messages[-1].content %}\n {%- endif %}\n {%- set reasoning_content = '' %}\n {%- if messages[-1].reasoning_content is string %}\n {%- set reasoning_content = messages[-1].reasoning_content %}\n {%- else %}\n {%- if '</think>' in content %}\n {%- set reasoning_content = content.split('</think>')[0].rstrip('\\n').split('<think>')[-1].lstrip('\\n') %}\n {%- set content = content.split('</think>')[-1].lstrip('\\n') %}\n {%- endif %}\n {%- endif %}\n {%- if reasoning_content %}\n{{- '<|im_start|>assistant\\n<think>\\n' + reasoning_content.strip('\\n') + '\\n</think>\\n\\n' + content.lstrip('\\n') + '<|im_end|>\\n' }}\n {%- else %}\n{{- '<|im_start|>assistant\\n' + content + '<|im_end|>\\n' }}\n {%- endif %}\n\n{%- elif messages and messages[-1].role == 'user' and messages[-1].content is string %}\n{{- '<|im_start|>user\\n' + messages[-1].content + '<|im_end|>\\n<|im_start|>assistant\\n' }}\n{%- endif %}"
241
- }
 
231
  "eos_token": "<|im_end|>",
232
  "errors": "replace",
233
  "extra_special_tokens": {},
234
+ "model_max_length": 262144,
235
+ "pad_token": "<|endoftext|>",
 
236
  "split_special_tokens": false,
237
  "tokenizer_class": "Qwen2Tokenizer",
238
+ "unk_token": null
239
+ }