Hugging Face's logo

RedHatAI
/
Llama-4-Maverick-17B-128E-Instruct-NVFP4

Text Generation
Safetensors
llama4
fp4
vllm
conversational
8-bit precision
compressed-tensors
Model card Files
xet
Community
1
Shubhra Pandit commited on
Commit
4bdd824
·
1 Parent(s): 54b1f91

Upload model files

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +2 -0
  2. chat_template.jinja +111 -0
  3. config.json +650 -0
  4. generation_config.json +13 -0
  5. model-00001-of-00047.safetensors +3 -0
  6. model-00002-of-00047.safetensors +3 -0
  7. model-00003-of-00047.safetensors +3 -0
  8. model-00004-of-00047.safetensors +3 -0
  9. model-00005-of-00047.safetensors +3 -0
  10. model-00006-of-00047.safetensors +3 -0
  11. model-00007-of-00047.safetensors +3 -0
  12. model-00008-of-00047.safetensors +3 -0
  13. model-00009-of-00047.safetensors +3 -0
  14. model-00010-of-00047.safetensors +3 -0
  15. model-00011-of-00047.safetensors +3 -0
  16. model-00012-of-00047.safetensors +3 -0
  17. model-00013-of-00047.safetensors +3 -0
  18. model-00014-of-00047.safetensors +3 -0
  19. model-00015-of-00047.safetensors +3 -0
  20. model-00016-of-00047.safetensors +3 -0
  21. model-00017-of-00047.safetensors +3 -0
  22. model-00018-of-00047.safetensors +3 -0
  23. model-00019-of-00047.safetensors +3 -0
  24. model-00020-of-00047.safetensors +3 -0
  25. model-00021-of-00047.safetensors +3 -0
  26. model-00022-of-00047.safetensors +3 -0
  27. model-00023-of-00047.safetensors +3 -0
  28. model-00024-of-00047.safetensors +3 -0
  29. model-00025-of-00047.safetensors +3 -0
  30. model-00026-of-00047.safetensors +3 -0
  31. model-00027-of-00047.safetensors +3 -0
  32. model-00028-of-00047.safetensors +3 -0
  33. model-00029-of-00047.safetensors +3 -0
  34. model-00030-of-00047.safetensors +3 -0
  35. model-00031-of-00047.safetensors +3 -0
  36. model-00032-of-00047.safetensors +3 -0
  37. model-00033-of-00047.safetensors +3 -0
  38. model-00034-of-00047.safetensors +3 -0
  39. model-00035-of-00047.safetensors +3 -0
  40. model-00036-of-00047.safetensors +3 -0
  41. model-00037-of-00047.safetensors +3 -0
  42. model-00038-of-00047.safetensors +3 -0
  43. model-00039-of-00047.safetensors +3 -0
  44. model-00040-of-00047.safetensors +3 -0
  45. model-00041-of-00047.safetensors +3 -0
  46. model-00042-of-00047.safetensors +3 -0
  47. model-00043-of-00047.safetensors +3 -0
  48. model-00044-of-00047.safetensors +3 -0
  49. model-00045-of-00047.safetensors +3 -0
  50. model-00046-of-00047.safetensors +3 -0
.gitattributes CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
37
+
chat_template.jinja ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {{- bos_token }}
2
+ {%- if custom_tools is defined and custom_tools%}
3
+ {%- set tools = custom_tools %}
4
+ {%- endif %}
5
+ {%- if tools is defined and tools %}
6
+ {%- set tool_definition = tool_definition ~ (tools | tojson(indent=4)) %}
7
+ {%- else %}
8
+ {%- set tools = none %}
9
+ {%- endif %}
10
+
11
+
12
+ {#- This block extracts the system message, so we can slot it into the right place. #}
13
+ {%- if messages[0]['role'] == 'system' %}
14
+ {%- set user_provided_system_message = true %}
15
+ {%- if messages[0]['content'] is string %}
16
+ {%- set system_message = messages[0]['content']|trim %}
17
+ {%- else %}
18
+ {%- set system_message = messages[0]['content'][0]['text']|trim %}
19
+ {%- endif %}
20
+ {%- set messages = messages[1:] %}
21
+ {%- else %}
22
+ {%- if tools is not none %}
23
+ {#- Since not system_message was provided by user, if tool is provided, system_message is now default tool system message #}
24
+ {#- This system message is from llama website:https://www.llama.com/docs/model-cards-and-prompt-formats/llama4/ #}
25
+ {%- set system_message = "You are a helpful assistant and an expert in function composition. You can answer general questions using your internal knowledge OR invoke functions when necessary. Follow these strict guidelines:\n\n1. FUNCTION CALLS:\n- ONLY use functions that are EXPLICITLY listed in the function list below\n- If NO functions are listed (empty function list []), respond ONLY with internal knowledge or \"I don't have access to [Unavailable service] information\"\n- If a function is not in the list, respond ONLY with internal knowledge or \"I don't have access to [Unavailable service] information\"\n- If ALL required parameters are present AND the query EXACTLY matches a listed function's purpose: output ONLY the function call(s)\n- Use exact format: [func_name1(param1=value1, param2=value2), func_name2(...)]\nExamples:\nCORRECT: [get_weather(location=\"Vancouver\"), calculate_route(start=\"Boston\", end=\"New York\")] <- Only if get_weather and calculate_route are in function list\nINCORRECT: get_weather(location=\"New York\")\nINCORRECT: Let me check the weather: [get_weather(location=\"New York\")]\nINCORRECT: [get_events(location=\"Singapore\")] <- If function not in list\n\n2. RESPONSE RULES:\n- For pure function requests matching a listed function: ONLY output the function call(s)\n- For knowledge questions: ONLY output text\n- For missing parameters: ONLY request the specific missing parameters\n- For unavailable services (not in function list): output ONLY with internal knowledge or \"I don't have access to [Unavailable service] information\". Do NOT execute a function call.\n- If the query asks for information beyond what a listed function provides: output ONLY with internal knowledge about your limitations\n- NEVER combine text and function calls in the same response\n- NEVER suggest alternative functions when the requested service is unavailable\n- NEVER create or invent new functions not listed below\n\n3. STRICT BOUNDARIES:\n- ONLY use functions from the list below - no exceptions\n- NEVER use a function as an alternative to unavailable information\n- NEVER call functions not present in the function list\n- NEVER add explanatory text to function calls\n- NEVER respond with empty brackets\n- Use proper Python/JSON syntax for function calls\n- Check the function list carefully before responding\n\n4. TOOL RESPONSE HANDLING:\n- When receiving tool responses: provide concise, natural language responses\n- Don't repeat tool response verbatim\n- Don't add supplementary information\n\nHere is a list of functions in JSON format that you can invoke:\n" %}
26
+ {%- else %}
27
+ {%- set system_message = "" %}
28
+ {%- endif %}
29
+ {%- endif %}
30
+ {#- Now writing the system message: use the user provided system message if user_provided_system_message, else default tool system message if tools presented #}
31
+ {%- if system_message %}
32
+ {#- always use user provided system message to override default tool system message #}
33
+ {{- "<|header_start|>system<|header_end|>\n\n" }}
34
+ {{- system_message }}
35
+ {%- if user_provided_system_message and tools %}
36
+ {{- "\nHere is a list of functions in JSON format that you can invoke. Use exact format: [func_name1(param1=value1, param2=value2), func_name2(...)]\n" }}
37
+ {{- tool_definition -}}
38
+ {%- elif tool_definition %}
39
+ {{- tool_definition -}}
40
+ {%- endif %}
41
+ {{- "<|eot|>" }}
42
+ {%- endif %}
43
+
44
+ {#- Now deal with all other messages #}
45
+ {%- for message in messages %}
46
+ {#- Base case: messages that are not from tool role and has empty tool_call list #}
47
+ {%- if not (message.role == 'ipython' or message.role == 'tool' or ('tool_calls' in message and message.tool_calls|length != 0 )) %}
48
+ {{- '<|header_start|>' + message['role'] + '<|header_end|>\n\n' }}
49
+ {%- if message['content'] is string %}
50
+ {{- message['content'] }}
51
+ {%- else %}
52
+ {%- for content in message['content'] %}
53
+ {%- if content['type'] == 'image' %}
54
+ {{- '<|image|>' }}
55
+ {%- elif content['type'] == 'text' %}
56
+ {{- content['text'] | trim }}
57
+ {%- endif %}
58
+ {%- endfor %}
59
+ {%- endif %}
60
+ {{- "<|eot|>" }}
61
+ {#- Tool case: messages has non-empty tool_call list, must from assistant #}
62
+ {%- elif 'tool_calls' in message %}
63
+ {#- assume tool_calls are always coming from assistant #}
64
+ {%- if message.role == 'assistant' %}
65
+ {{- '<|header_start|>assistant<|header_end|>\n\n' -}}
66
+ {%- if message['content'] is string %}
67
+ {{- message['content'] }}
68
+ {%- else %}
69
+ {%- for content in message['content'] %}
70
+ {%- if content['type'] == 'image' %}
71
+ {{- '<|image|>' }}
72
+ {%- elif content['type'] == 'text' %}
73
+ {{- content['text'] }}
74
+ {%- endif %}
75
+ {%- endfor %}
76
+ {%- endif %}
77
+ {{- "[" }}
78
+ {%- for tool_call in message.tool_calls %}
79
+ {%- if tool_call.function is defined %}
80
+ {%- set tool_call = tool_call.function %}
81
+ {%- endif %}
82
+ {{- tool_call.name + '(' -}}
83
+ {%- for param in tool_call.arguments %}
84
+ {{- param + '="' -}}
85
+ {{- "%s" | format(tool_call.arguments[param]) -}}
86
+ {{- '"' -}}
87
+ {% if not loop.last %}, {% endif %}
88
+ {%- endfor %}
89
+ {{- ')' -}}
90
+ {% if not loop.last %}, {% endif %}
91
+ {%- endfor %}
92
+ {{- "]<|eot|>" }}
93
+ {%- endif %}
94
+ {#- Tool_response case: messages are from tool_response #}
95
+ {%- elif message.role == "tool" or message.role == "ipython" %}
96
+ {{- "<|header_start|>ipython<|header_end|>\n\n" }}
97
+ {%- if message.content is string %}
98
+ {{- message.content | tojson }}
99
+ {%- else %}
100
+ {%- for content in message['content'] %}
101
+ {%- if content['type'] == 'text' %}
102
+ {{- content['text'] | tojson }}
103
+ {%- endif %}
104
+ {%- endfor %}
105
+ {%- endif %}
106
+ {{- "<|eot|>" }}
107
+ {%- endif %}
108
+ {%- endfor %}
109
+ {%- if add_generation_prompt %}
110
+ {{- '<|header_start|>assistant<|header_end|>\n\n' }}
111
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,650 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Llama4ForConditionalGeneration"
4
+ ],
5
+ "boi_token_index": 200080,
6
+ "dtype": "bfloat16",
7
+ "eoi_token_index": 200081,
8
+ "image_token_index": 200092,
9
+ "model_type": "llama4",
10
+ "quantization_config": {
11
+ "config_groups": {
12
+ "group_0": {
13
+ "format": "nvfp4-pack-quantized",
14
+ "input_activations": {
15
+ "actorder": null,
16
+ "block_structure": null,
17
+ "dynamic": "local",
18
+ "group_size": 16,
19
+ "num_bits": 4,
20
+ "observer": "minmax",
21
+ "observer_kwargs": {},
22
+ "strategy": "tensor_group",
23
+ "symmetric": true,
24
+ "type": "float"
25
+ },
26
+ "output_activations": null,
27
+ "targets": [
28
+ "Linear"
29
+ ],
30
+ "weights": {
31
+ "actorder": null,
32
+ "block_structure": null,
33
+ "dynamic": false,
34
+ "group_size": 16,
35
+ "num_bits": 4,
36
+ "observer": "minmax",
37
+ "observer_kwargs": {},
38
+ "strategy": "tensor_group",
39
+ "symmetric": true,
40
+ "type": "float"
41
+ }
42
+ }
43
+ },
44
+ "format": "nvfp4-pack-quantized",
45
+ "global_compression_ratio": null,
46
+ "ignore": [
47
+ "vision_model.patch_embedding.linear",
48
+ "vision_model.model.layers.0.self_attn.q_proj",
49
+ "vision_model.model.layers.0.self_attn.k_proj",
50
+ "vision_model.model.layers.0.self_attn.v_proj",
51
+ "vision_model.model.layers.0.self_attn.o_proj",
52
+ "vision_model.model.layers.0.mlp.fc1",
53
+ "vision_model.model.layers.0.mlp.fc2",
54
+ "vision_model.model.layers.1.self_attn.q_proj",
55
+ "vision_model.model.layers.1.self_attn.k_proj",
56
+ "vision_model.model.layers.1.self_attn.v_proj",
57
+ "vision_model.model.layers.1.self_attn.o_proj",
58
+ "vision_model.model.layers.1.mlp.fc1",
59
+ "vision_model.model.layers.1.mlp.fc2",
60
+ "vision_model.model.layers.2.self_attn.q_proj",
61
+ "vision_model.model.layers.2.self_attn.k_proj",
62
+ "vision_model.model.layers.2.self_attn.v_proj",
63
+ "vision_model.model.layers.2.self_attn.o_proj",
64
+ "vision_model.model.layers.2.mlp.fc1",
65
+ "vision_model.model.layers.2.mlp.fc2",
66
+ "vision_model.model.layers.3.self_attn.q_proj",
67
+ "vision_model.model.layers.3.self_attn.k_proj",
68
+ "vision_model.model.layers.3.self_attn.v_proj",
69
+ "vision_model.model.layers.3.self_attn.o_proj",
70
+ "vision_model.model.layers.3.mlp.fc1",
71
+ "vision_model.model.layers.3.mlp.fc2",
72
+ "vision_model.model.layers.4.self_attn.q_proj",
73
+ "vision_model.model.layers.4.self_attn.k_proj",
74
+ "vision_model.model.layers.4.self_attn.v_proj",
75
+ "vision_model.model.layers.4.self_attn.o_proj",
76
+ "vision_model.model.layers.4.mlp.fc1",
77
+ "vision_model.model.layers.4.mlp.fc2",
78
+ "vision_model.model.layers.5.self_attn.q_proj",
79
+ "vision_model.model.layers.5.self_attn.k_proj",
80
+ "vision_model.model.layers.5.self_attn.v_proj",
81
+ "vision_model.model.layers.5.self_attn.o_proj",
82
+ "vision_model.model.layers.5.mlp.fc1",
83
+ "vision_model.model.layers.5.mlp.fc2",
84
+ "vision_model.model.layers.6.self_attn.q_proj",
85
+ "vision_model.model.layers.6.self_attn.k_proj",
86
+ "vision_model.model.layers.6.self_attn.v_proj",
87
+ "vision_model.model.layers.6.self_attn.o_proj",
88
+ "vision_model.model.layers.6.mlp.fc1",
89
+ "vision_model.model.layers.6.mlp.fc2",
90
+ "vision_model.model.layers.7.self_attn.q_proj",
91
+ "vision_model.model.layers.7.self_attn.k_proj",
92
+ "vision_model.model.layers.7.self_attn.v_proj",
93
+ "vision_model.model.layers.7.self_attn.o_proj",
94
+ "vision_model.model.layers.7.mlp.fc1",
95
+ "vision_model.model.layers.7.mlp.fc2",
96
+ "vision_model.model.layers.8.self_attn.q_proj",
97
+ "vision_model.model.layers.8.self_attn.k_proj",
98
+ "vision_model.model.layers.8.self_attn.v_proj",
99
+ "vision_model.model.layers.8.self_attn.o_proj",
100
+ "vision_model.model.layers.8.mlp.fc1",
101
+ "vision_model.model.layers.8.mlp.fc2",
102
+ "vision_model.model.layers.9.self_attn.q_proj",
103
+ "vision_model.model.layers.9.self_attn.k_proj",
104
+ "vision_model.model.layers.9.self_attn.v_proj",
105
+ "vision_model.model.layers.9.self_attn.o_proj",
106
+ "vision_model.model.layers.9.mlp.fc1",
107
+ "vision_model.model.layers.9.mlp.fc2",
108
+ "vision_model.model.layers.10.self_attn.q_proj",
109
+ "vision_model.model.layers.10.self_attn.k_proj",
110
+ "vision_model.model.layers.10.self_attn.v_proj",
111
+ "vision_model.model.layers.10.self_attn.o_proj",
112
+ "vision_model.model.layers.10.mlp.fc1",
113
+ "vision_model.model.layers.10.mlp.fc2",
114
+ "vision_model.model.layers.11.self_attn.q_proj",
115
+ "vision_model.model.layers.11.self_attn.k_proj",
116
+ "vision_model.model.layers.11.self_attn.v_proj",
117
+ "vision_model.model.layers.11.self_attn.o_proj",
118
+ "vision_model.model.layers.11.mlp.fc1",
119
+ "vision_model.model.layers.11.mlp.fc2",
120
+ "vision_model.model.layers.12.self_attn.q_proj",
121
+ "vision_model.model.layers.12.self_attn.k_proj",
122
+ "vision_model.model.layers.12.self_attn.v_proj",
123
+ "vision_model.model.layers.12.self_attn.o_proj",
124
+ "vision_model.model.layers.12.mlp.fc1",
125
+ "vision_model.model.layers.12.mlp.fc2",
126
+ "vision_model.model.layers.13.self_attn.q_proj",
127
+ "vision_model.model.layers.13.self_attn.k_proj",
128
+ "vision_model.model.layers.13.self_attn.v_proj",
129
+ "vision_model.model.layers.13.self_attn.o_proj",
130
+ "vision_model.model.layers.13.mlp.fc1",
131
+ "vision_model.model.layers.13.mlp.fc2",
132
+ "vision_model.model.layers.14.self_attn.q_proj",
133
+ "vision_model.model.layers.14.self_attn.k_proj",
134
+ "vision_model.model.layers.14.self_attn.v_proj",
135
+ "vision_model.model.layers.14.self_attn.o_proj",
136
+ "vision_model.model.layers.14.mlp.fc1",
137
+ "vision_model.model.layers.14.mlp.fc2",
138
+ "vision_model.model.layers.15.self_attn.q_proj",
139
+ "vision_model.model.layers.15.self_attn.k_proj",
140
+ "vision_model.model.layers.15.self_attn.v_proj",
141
+ "vision_model.model.layers.15.self_attn.o_proj",
142
+ "vision_model.model.layers.15.mlp.fc1",
143
+ "vision_model.model.layers.15.mlp.fc2",
144
+ "vision_model.model.layers.16.self_attn.q_proj",
145
+ "vision_model.model.layers.16.self_attn.k_proj",
146
+ "vision_model.model.layers.16.self_attn.v_proj",
147
+ "vision_model.model.layers.16.self_attn.o_proj",
148
+ "vision_model.model.layers.16.mlp.fc1",
149
+ "vision_model.model.layers.16.mlp.fc2",
150
+ "vision_model.model.layers.17.self_attn.q_proj",
151
+ "vision_model.model.layers.17.self_attn.k_proj",
152
+ "vision_model.model.layers.17.self_attn.v_proj",
153
+ "vision_model.model.layers.17.self_attn.o_proj",
154
+ "vision_model.model.layers.17.mlp.fc1",
155
+ "vision_model.model.layers.17.mlp.fc2",
156
+ "vision_model.model.layers.18.self_attn.q_proj",
157
+ "vision_model.model.layers.18.self_attn.k_proj",
158
+ "vision_model.model.layers.18.self_attn.v_proj",
159
+ "vision_model.model.layers.18.self_attn.o_proj",
160
+ "vision_model.model.layers.18.mlp.fc1",
161
+ "vision_model.model.layers.18.mlp.fc2",
162
+ "vision_model.model.layers.19.self_attn.q_proj",
163
+ "vision_model.model.layers.19.self_attn.k_proj",
164
+ "vision_model.model.layers.19.self_attn.v_proj",
165
+ "vision_model.model.layers.19.self_attn.o_proj",
166
+ "vision_model.model.layers.19.mlp.fc1",
167
+ "vision_model.model.layers.19.mlp.fc2",
168
+ "vision_model.model.layers.20.self_attn.q_proj",
169
+ "vision_model.model.layers.20.self_attn.k_proj",
170
+ "vision_model.model.layers.20.self_attn.v_proj",
171
+ "vision_model.model.layers.20.self_attn.o_proj",
172
+ "vision_model.model.layers.20.mlp.fc1",
173
+ "vision_model.model.layers.20.mlp.fc2",
174
+ "vision_model.model.layers.21.self_attn.q_proj",
175
+ "vision_model.model.layers.21.self_attn.k_proj",
176
+ "vision_model.model.layers.21.self_attn.v_proj",
177
+ "vision_model.model.layers.21.self_attn.o_proj",
178
+ "vision_model.model.layers.21.mlp.fc1",
179
+ "vision_model.model.layers.21.mlp.fc2",
180
+ "vision_model.model.layers.22.self_attn.q_proj",
181
+ "vision_model.model.layers.22.self_attn.k_proj",
182
+ "vision_model.model.layers.22.self_attn.v_proj",
183
+ "vision_model.model.layers.22.self_attn.o_proj",
184
+ "vision_model.model.layers.22.mlp.fc1",
185
+ "vision_model.model.layers.22.mlp.fc2",
186
+ "vision_model.model.layers.23.self_attn.q_proj",
187
+ "vision_model.model.layers.23.self_attn.k_proj",
188
+ "vision_model.model.layers.23.self_attn.v_proj",
189
+ "vision_model.model.layers.23.self_attn.o_proj",
190
+ "vision_model.model.layers.23.mlp.fc1",
191
+ "vision_model.model.layers.23.mlp.fc2",
192
+ "vision_model.model.layers.24.self_attn.q_proj",
193
+ "vision_model.model.layers.24.self_attn.k_proj",
194
+ "vision_model.model.layers.24.self_attn.v_proj",
195
+ "vision_model.model.layers.24.self_attn.o_proj",
196
+ "vision_model.model.layers.24.mlp.fc1",
197
+ "vision_model.model.layers.24.mlp.fc2",
198
+ "vision_model.model.layers.25.self_attn.q_proj",
199
+ "vision_model.model.layers.25.self_attn.k_proj",
200
+ "vision_model.model.layers.25.self_attn.v_proj",
201
+ "vision_model.model.layers.25.self_attn.o_proj",
202
+ "vision_model.model.layers.25.mlp.fc1",
203
+ "vision_model.model.layers.25.mlp.fc2",
204
+ "vision_model.model.layers.26.self_attn.q_proj",
205
+ "vision_model.model.layers.26.self_attn.k_proj",
206
+ "vision_model.model.layers.26.self_attn.v_proj",
207
+ "vision_model.model.layers.26.self_attn.o_proj",
208
+ "vision_model.model.layers.26.mlp.fc1",
209
+ "vision_model.model.layers.26.mlp.fc2",
210
+ "vision_model.model.layers.27.self_attn.q_proj",
211
+ "vision_model.model.layers.27.self_attn.k_proj",
212
+ "vision_model.model.layers.27.self_attn.v_proj",
213
+ "vision_model.model.layers.27.self_attn.o_proj",
214
+ "vision_model.model.layers.27.mlp.fc1",
215
+ "vision_model.model.layers.27.mlp.fc2",
216
+ "vision_model.model.layers.28.self_attn.q_proj",
217
+ "vision_model.model.layers.28.self_attn.k_proj",
218
+ "vision_model.model.layers.28.self_attn.v_proj",
219
+ "vision_model.model.layers.28.self_attn.o_proj",
220
+ "vision_model.model.layers.28.mlp.fc1",
221
+ "vision_model.model.layers.28.mlp.fc2",
222
+ "vision_model.model.layers.29.self_attn.q_proj",
223
+ "vision_model.model.layers.29.self_attn.k_proj",
224
+ "vision_model.model.layers.29.self_attn.v_proj",
225
+ "vision_model.model.layers.29.self_attn.o_proj",
226
+ "vision_model.model.layers.29.mlp.fc1",
227
+ "vision_model.model.layers.29.mlp.fc2",
228
+ "vision_model.model.layers.30.self_attn.q_proj",
229
+ "vision_model.model.layers.30.self_attn.k_proj",
230
+ "vision_model.model.layers.30.self_attn.v_proj",
231
+ "vision_model.model.layers.30.self_attn.o_proj",
232
+ "vision_model.model.layers.30.mlp.fc1",
233
+ "vision_model.model.layers.30.mlp.fc2",
234
+ "vision_model.model.layers.31.self_attn.q_proj",
235
+ "vision_model.model.layers.31.self_attn.k_proj",
236
+ "vision_model.model.layers.31.self_attn.v_proj",
237
+ "vision_model.model.layers.31.self_attn.o_proj",
238
+ "vision_model.model.layers.31.mlp.fc1",
239
+ "vision_model.model.layers.31.mlp.fc2",
240
+ "vision_model.model.layers.32.self_attn.q_proj",
241
+ "vision_model.model.layers.32.self_attn.k_proj",
242
+ "vision_model.model.layers.32.self_attn.v_proj",
243
+ "vision_model.model.layers.32.self_attn.o_proj",
244
+ "vision_model.model.layers.32.mlp.fc1",
245
+ "vision_model.model.layers.32.mlp.fc2",
246
+ "vision_model.model.layers.33.self_attn.q_proj",
247
+ "vision_model.model.layers.33.self_attn.k_proj",
248
+ "vision_model.model.layers.33.self_attn.v_proj",
249
+ "vision_model.model.layers.33.self_attn.o_proj",
250
+ "vision_model.model.layers.33.mlp.fc1",
251
+ "vision_model.model.layers.33.mlp.fc2",
252
+ "vision_model.vision_adapter.mlp.fc1",
253
+ "vision_model.vision_adapter.mlp.fc2",
254
+ "multi_modal_projector.linear_1",
255
+ "language_model.model.layers.0.self_attn.q_proj",
256
+ "language_model.model.layers.0.self_attn.k_proj",
257
+ "language_model.model.layers.0.self_attn.v_proj",
258
+ "language_model.model.layers.0.self_attn.o_proj",
259
+ "language_model.model.layers.1.self_attn.q_proj",
260
+ "language_model.model.layers.1.self_attn.k_proj",
261
+ "language_model.model.layers.1.self_attn.v_proj",
262
+ "language_model.model.layers.1.self_attn.o_proj",
263
+ "language_model.model.layers.2.self_attn.q_proj",
264
+ "language_model.model.layers.2.self_attn.k_proj",
265
+ "language_model.model.layers.2.self_attn.v_proj",
266
+ "language_model.model.layers.2.self_attn.o_proj",
267
+ "language_model.model.layers.3.self_attn.q_proj",
268
+ "language_model.model.layers.3.self_attn.k_proj",
269
+ "language_model.model.layers.3.self_attn.v_proj",
270
+ "language_model.model.layers.3.self_attn.o_proj",
271
+ "language_model.model.layers.4.self_attn.q_proj",
272
+ "language_model.model.layers.4.self_attn.k_proj",
273
+ "language_model.model.layers.4.self_attn.v_proj",
274
+ "language_model.model.layers.4.self_attn.o_proj",
275
+ "language_model.model.layers.5.self_attn.q_proj",
276
+ "language_model.model.layers.5.self_attn.k_proj",
277
+ "language_model.model.layers.5.self_attn.v_proj",
278
+ "language_model.model.layers.5.self_attn.o_proj",
279
+ "language_model.model.layers.6.self_attn.q_proj",
280
+ "language_model.model.layers.6.self_attn.k_proj",
281
+ "language_model.model.layers.6.self_attn.v_proj",
282
+ "language_model.model.layers.6.self_attn.o_proj",
283
+ "language_model.model.layers.7.self_attn.q_proj",
284
+ "language_model.model.layers.7.self_attn.k_proj",
285
+ "language_model.model.layers.7.self_attn.v_proj",
286
+ "language_model.model.layers.7.self_attn.o_proj",
287
+ "language_model.model.layers.8.self_attn.q_proj",
288
+ "language_model.model.layers.8.self_attn.k_proj",
289
+ "language_model.model.layers.8.self_attn.v_proj",
290
+ "language_model.model.layers.8.self_attn.o_proj",
291
+ "language_model.model.layers.9.self_attn.q_proj",
292
+ "language_model.model.layers.9.self_attn.k_proj",
293
+ "language_model.model.layers.9.self_attn.v_proj",
294
+ "language_model.model.layers.9.self_attn.o_proj",
295
+ "language_model.model.layers.10.self_attn.q_proj",
296
+ "language_model.model.layers.10.self_attn.k_proj",
297
+ "language_model.model.layers.10.self_attn.v_proj",
298
+ "language_model.model.layers.10.self_attn.o_proj",
299
+ "language_model.model.layers.11.self_attn.q_proj",
300
+ "language_model.model.layers.11.self_attn.k_proj",
301
+ "language_model.model.layers.11.self_attn.v_proj",
302
+ "language_model.model.layers.11.self_attn.o_proj",
303
+ "language_model.model.layers.12.self_attn.q_proj",
304
+ "language_model.model.layers.12.self_attn.k_proj",
305
+ "language_model.model.layers.12.self_attn.v_proj",
306
+ "language_model.model.layers.12.self_attn.o_proj",
307
+ "language_model.model.layers.13.self_attn.q_proj",
308
+ "language_model.model.layers.13.self_attn.k_proj",
309
+ "language_model.model.layers.13.self_attn.v_proj",
310
+ "language_model.model.layers.13.self_attn.o_proj",
311
+ "language_model.model.layers.14.self_attn.q_proj",
312
+ "language_model.model.layers.14.self_attn.k_proj",
313
+ "language_model.model.layers.14.self_attn.v_proj",
314
+ "language_model.model.layers.14.self_attn.o_proj",
315
+ "language_model.model.layers.15.self_attn.q_proj",
316
+ "language_model.model.layers.15.self_attn.k_proj",
317
+ "language_model.model.layers.15.self_attn.v_proj",
318
+ "language_model.model.layers.15.self_attn.o_proj",
319
+ "language_model.model.layers.16.self_attn.q_proj",
320
+ "language_model.model.layers.16.self_attn.k_proj",
321
+ "language_model.model.layers.16.self_attn.v_proj",
322
+ "language_model.model.layers.16.self_attn.o_proj",
323
+ "language_model.model.layers.17.self_attn.q_proj",
324
+ "language_model.model.layers.17.self_attn.k_proj",
325
+ "language_model.model.layers.17.self_attn.v_proj",
326
+ "language_model.model.layers.17.self_attn.o_proj",
327
+ "language_model.model.layers.18.self_attn.q_proj",
328
+ "language_model.model.layers.18.self_attn.k_proj",
329
+ "language_model.model.layers.18.self_attn.v_proj",
330
+ "language_model.model.layers.18.self_attn.o_proj",
331
+ "language_model.model.layers.19.self_attn.q_proj",
332
+ "language_model.model.layers.19.self_attn.k_proj",
333
+ "language_model.model.layers.19.self_attn.v_proj",
334
+ "language_model.model.layers.19.self_attn.o_proj",
335
+ "language_model.model.layers.20.self_attn.q_proj",
336
+ "language_model.model.layers.20.self_attn.k_proj",
337
+ "language_model.model.layers.20.self_attn.v_proj",
338
+ "language_model.model.layers.20.self_attn.o_proj",
339
+ "language_model.model.layers.21.self_attn.q_proj",
340
+ "language_model.model.layers.21.self_attn.k_proj",
341
+ "language_model.model.layers.21.self_attn.v_proj",
342
+ "language_model.model.layers.21.self_attn.o_proj",
343
+ "language_model.model.layers.22.self_attn.q_proj",
344
+ "language_model.model.layers.22.self_attn.k_proj",
345
+ "language_model.model.layers.22.self_attn.v_proj",
346
+ "language_model.model.layers.22.self_attn.o_proj",
347
+ "language_model.model.layers.23.self_attn.q_proj",
348
+ "language_model.model.layers.23.self_attn.k_proj",
349
+ "language_model.model.layers.23.self_attn.v_proj",
350
+ "language_model.model.layers.23.self_attn.o_proj",
351
+ "language_model.model.layers.24.self_attn.q_proj",
352
+ "language_model.model.layers.24.self_attn.k_proj",
353
+ "language_model.model.layers.24.self_attn.v_proj",
354
+ "language_model.model.layers.24.self_attn.o_proj",
355
+ "language_model.model.layers.25.self_attn.q_proj",
356
+ "language_model.model.layers.25.self_attn.k_proj",
357
+ "language_model.model.layers.25.self_attn.v_proj",
358
+ "language_model.model.layers.25.self_attn.o_proj",
359
+ "language_model.model.layers.26.self_attn.q_proj",
360
+ "language_model.model.layers.26.self_attn.k_proj",
361
+ "language_model.model.layers.26.self_attn.v_proj",
362
+ "language_model.model.layers.26.self_attn.o_proj",
363
+ "language_model.model.layers.27.self_attn.q_proj",
364
+ "language_model.model.layers.27.self_attn.k_proj",
365
+ "language_model.model.layers.27.self_attn.v_proj",
366
+ "language_model.model.layers.27.self_attn.o_proj",
367
+ "language_model.model.layers.28.self_attn.q_proj",
368
+ "language_model.model.layers.28.self_attn.k_proj",
369
+ "language_model.model.layers.28.self_attn.v_proj",
370
+ "language_model.model.layers.28.self_attn.o_proj",
371
+ "language_model.model.layers.29.self_attn.q_proj",
372
+ "language_model.model.layers.29.self_attn.k_proj",
373
+ "language_model.model.layers.29.self_attn.v_proj",
374
+ "language_model.model.layers.29.self_attn.o_proj",
375
+ "language_model.model.layers.30.self_attn.q_proj",
376
+ "language_model.model.layers.30.self_attn.k_proj",
377
+ "language_model.model.layers.30.self_attn.v_proj",
378
+ "language_model.model.layers.30.self_attn.o_proj",
379
+ "language_model.model.layers.31.self_attn.q_proj",
380
+ "language_model.model.layers.31.self_attn.k_proj",
381
+ "language_model.model.layers.31.self_attn.v_proj",
382
+ "language_model.model.layers.31.self_attn.o_proj",
383
+ "language_model.model.layers.32.self_attn.q_proj",
384
+ "language_model.model.layers.32.self_attn.k_proj",
385
+ "language_model.model.layers.32.self_attn.v_proj",
386
+ "language_model.model.layers.32.self_attn.o_proj",
387
+ "language_model.model.layers.33.self_attn.q_proj",
388
+ "language_model.model.layers.33.self_attn.k_proj",
389
+ "language_model.model.layers.33.self_attn.v_proj",
390
+ "language_model.model.layers.33.self_attn.o_proj",
391
+ "language_model.model.layers.34.self_attn.q_proj",
392
+ "language_model.model.layers.34.self_attn.k_proj",
393
+ "language_model.model.layers.34.self_attn.v_proj",
394
+ "language_model.model.layers.34.self_attn.o_proj",
395
+ "language_model.model.layers.35.self_attn.q_proj",
396
+ "language_model.model.layers.35.self_attn.k_proj",
397
+ "language_model.model.layers.35.self_attn.v_proj",
398
+ "language_model.model.layers.35.self_attn.o_proj",
399
+ "language_model.model.layers.36.self_attn.q_proj",
400
+ "language_model.model.layers.36.self_attn.k_proj",
401
+ "language_model.model.layers.36.self_attn.v_proj",
402
+ "language_model.model.layers.36.self_attn.o_proj",
403
+ "language_model.model.layers.37.self_attn.q_proj",
404
+ "language_model.model.layers.37.self_attn.k_proj",
405
+ "language_model.model.layers.37.self_attn.v_proj",
406
+ "language_model.model.layers.37.self_attn.o_proj",
407
+ "language_model.model.layers.38.self_attn.q_proj",
408
+ "language_model.model.layers.38.self_attn.k_proj",
409
+ "language_model.model.layers.38.self_attn.v_proj",
410
+ "language_model.model.layers.38.self_attn.o_proj",
411
+ "language_model.model.layers.39.self_attn.q_proj",
412
+ "language_model.model.layers.39.self_attn.k_proj",
413
+ "language_model.model.layers.39.self_attn.v_proj",
414
+ "language_model.model.layers.39.self_attn.o_proj",
415
+ "language_model.model.layers.40.self_attn.q_proj",
416
+ "language_model.model.layers.40.self_attn.k_proj",
417
+ "language_model.model.layers.40.self_attn.v_proj",
418
+ "language_model.model.layers.40.self_attn.o_proj",
419
+ "language_model.model.layers.41.self_attn.q_proj",
420
+ "language_model.model.layers.41.self_attn.k_proj",
421
+ "language_model.model.layers.41.self_attn.v_proj",
422
+ "language_model.model.layers.41.self_attn.o_proj",
423
+ "language_model.model.layers.42.self_attn.q_proj",
424
+ "language_model.model.layers.42.self_attn.k_proj",
425
+ "language_model.model.layers.42.self_attn.v_proj",
426
+ "language_model.model.layers.42.self_attn.o_proj",
427
+ "language_model.model.layers.43.self_attn.q_proj",
428
+ "language_model.model.layers.43.self_attn.k_proj",
429
+ "language_model.model.layers.43.self_attn.v_proj",
430
+ "language_model.model.layers.43.self_attn.o_proj",
431
+ "language_model.model.layers.44.self_attn.q_proj",
432
+ "language_model.model.layers.44.self_attn.k_proj",
433
+ "language_model.model.layers.44.self_attn.v_proj",
434
+ "language_model.model.layers.44.self_attn.o_proj",
435
+ "language_model.model.layers.45.self_attn.q_proj",
436
+ "language_model.model.layers.45.self_attn.k_proj",
437
+ "language_model.model.layers.45.self_attn.v_proj",
438
+ "language_model.model.layers.45.self_attn.o_proj",
439
+ "language_model.model.layers.46.self_attn.q_proj",
440
+ "language_model.model.layers.46.self_attn.k_proj",
441
+ "language_model.model.layers.46.self_attn.v_proj",
442
+ "language_model.model.layers.46.self_attn.o_proj",
443
+ "language_model.model.layers.47.self_attn.q_proj",
444
+ "language_model.model.layers.47.self_attn.k_proj",
445
+ "language_model.model.layers.47.self_attn.v_proj",
446
+ "language_model.model.layers.47.self_attn.o_proj",
447
+ "language_model.lm_head"
448
+ ],
449
+ "kv_cache_scheme": null,
450
+ "quant_method": "compressed-tensors",
451
+ "quantization_status": "compressed",
452
+ "sparsity_config": {},
453
+ "transform_config": {},
454
+ "version": "0.12.3.a20251009"
455
+ },
456
+ "text_config": {
457
+ "_attn_implementation_autoset": true,
458
+ "attention_bias": false,
459
+ "attention_chunk_size": 8192,
460
+ "attention_dropout": 0.0,
461
+ "attn_scale": 0.1,
462
+ "attn_temperature_tuning": true,
463
+ "bos_token_id": 200000,
464
+ "dtype": "bfloat16",
465
+ "eos_token_id": [
466
+ 200001,
467
+ 200007,
468
+ 200008
469
+ ],
470
+ "floor_scale": 8192,
471
+ "for_llm_compressor": false,
472
+ "head_dim": 128,
473
+ "hidden_act": "silu",
474
+ "hidden_size": 5120,
475
+ "initializer_range": 0.02,
476
+ "interleave_moe_layer_step": 2,
477
+ "intermediate_size": 8192,
478
+ "intermediate_size_mlp": 16384,
479
+ "layer_types": [
480
+ "chunked_attention",
481
+ "chunked_attention",
482
+ "chunked_attention",
483
+ "full_attention",
484
+ "chunked_attention",
485
+ "chunked_attention",
486
+ "chunked_attention",
487
+ "full_attention",
488
+ "chunked_attention",
489
+ "chunked_attention",
490
+ "chunked_attention",
491
+ "full_attention",
492
+ "chunked_attention",
493
+ "chunked_attention",
494
+ "chunked_attention",
495
+ "full_attention",
496
+ "chunked_attention",
497
+ "chunked_attention",
498
+ "chunked_attention",
499
+ "full_attention",
500
+ "chunked_attention",
501
+ "chunked_attention",
502
+ "chunked_attention",
503
+ "full_attention",
504
+ "chunked_attention",
505
+ "chunked_attention",
506
+ "chunked_attention",
507
+ "full_attention",
508
+ "chunked_attention",
509
+ "chunked_attention",
510
+ "chunked_attention",
511
+ "full_attention",
512
+ "chunked_attention",
513
+ "chunked_attention",
514
+ "chunked_attention",
515
+ "full_attention",
516
+ "chunked_attention",
517
+ "chunked_attention",
518
+ "chunked_attention",
519
+ "full_attention",
520
+ "chunked_attention",
521
+ "chunked_attention",
522
+ "chunked_attention",
523
+ "full_attention",
524
+ "chunked_attention",
525
+ "chunked_attention",
526
+ "chunked_attention",
527
+ "full_attention"
528
+ ],
529
+ "max_position_embeddings": 1048576,
530
+ "model_type": "llama4_text",
531
+ "moe_layers": [
532
+ 1,
533
+ 3,
534
+ 5,
535
+ 7,
536
+ 9,
537
+ 11,
538
+ 13,
539
+ 15,
540
+ 17,
541
+ 19,
542
+ 21,
543
+ 23,
544
+ 25,
545
+ 27,
546
+ 29,
547
+ 31,
548
+ 33,
549
+ 35,
550
+ 37,
551
+ 39,
552
+ 41,
553
+ 43,
554
+ 45,
555
+ 47
556
+ ],
557
+ "no_rope_layers": [
558
+ 1,
559
+ 1,
560
+ 1,
561
+ 0,
562
+ 1,
563
+ 1,
564
+ 1,
565
+ 0,
566
+ 1,
567
+ 1,
568
+ 1,
569
+ 0,
570
+ 1,
571
+ 1,
572
+ 1,
573
+ 0,
574
+ 1,
575
+ 1,
576
+ 1,
577
+ 0,
578
+ 1,
579
+ 1,
580
+ 1,
581
+ 0,
582
+ 1,
583
+ 1,
584
+ 1,
585
+ 0,
586
+ 1,
587
+ 1,
588
+ 1,
589
+ 0,
590
+ 1,
591
+ 1,
592
+ 1,
593
+ 0,
594
+ 1,
595
+ 1,
596
+ 1,
597
+ 0,
598
+ 1,
599
+ 1,
600
+ 1,
601
+ 0,
602
+ 1,
603
+ 1,
604
+ 1,
605
+ 0
606
+ ],
607
+ "num_attention_heads": 40,
608
+ "num_experts_per_tok": 1,
609
+ "num_hidden_layers": 48,
610
+ "num_key_value_heads": 8,
611
+ "num_local_experts": 128,
612
+ "output_router_logits": false,
613
+ "pad_token_id": 200018,
614
+ "rms_norm_eps": 1e-05,
615
+ "rope_scaling": null,
616
+ "rope_theta": 500000.0,
617
+ "router_aux_loss_coef": 0.001,
618
+ "router_jitter_noise": 0.0,
619
+ "use_cache": true,
620
+ "use_qk_norm": false,
621
+ "vocab_size": 202048
622
+ },
623
+ "tie_word_embeddings": false,
624
+ "transformers_version": "4.57.0",
625
+ "vision_config": {
626
+ "_attn_implementation_autoset": true,
627
+ "_vision_feature_layer": -1,
628
+ "attention_dropout": 0.0,
629
+ "hidden_act": "gelu",
630
+ "hidden_size": 1408,
631
+ "image_size": 336,
632
+ "initializer_range": 0.02,
633
+ "intermediate_size": 5632,
634
+ "model_type": "llama4_vision_model",
635
+ "multi_modal_projector_bias": false,
636
+ "norm_eps": 1e-05,
637
+ "num_attention_heads": 16,
638
+ "num_channels": 3,
639
+ "num_hidden_layers": 34,
640
+ "patch_size": 14,
641
+ "pixel_shuffle_ratio": 0.5,
642
+ "projector_dropout": 0.0,
643
+ "projector_input_dim": 4096,
644
+ "projector_output_dim": 4096,
645
+ "rope_theta": 10000,
646
+ "vision_feature_layer": -1,
647
+ "vision_feature_select_strategy": "default",
648
+ "vision_output_dim": 4096
649
+ }
650
+ }
generation_config.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 200000,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 200001,
6
+ 200007,
7
+ 200008
8
+ ],
9
+ "pad_token_id": 200018,
10
+ "temperature": 0.6,
11
+ "top_p": 0.9,
12
+ "transformers_version": "4.57.0"
13
+ }
model-00001-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:47b040ac26ba79642d8054af8f8d3cb1b7c40b0085aabfc41e14f448225fc744
3
+ size 4984732280
model-00002-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9576752e6cb75a93e640ca8f4f03346c588e3d1e6365367cdd01cadd2fbfc964
3
+ size 4999211496
model-00003-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9efef93b1661f6e7283a0a8ccf0e934a78e10eca6981c87cfa472763441f2c6c
3
+ size 4997935600
model-00004-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0e808e961066016325ee86ca19bb8bb8f5dc03c9f8879e70996d1d7fd32fc7e
3
+ size 4999211504
model-00005-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:11580d4d533020208b687b415fef05ea834ff5458e0b0c6f9719d77bd647b843
3
+ size 4997935600
model-00006-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57f10646eb2eff2771a9aa83b31936554e0af81723bb0640a067b0addcaa24d5
3
+ size 4999211496
model-00007-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:db4494d9ef86d92c70364cc8bb85650674b0179806d17d4a7ed346a5b337aa0f
3
+ size 4997935600
model-00008-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4da73a02893fc52ab99b69ddaaf4d08bf19fdb350a08712825569fde35feeca1
3
+ size 4999211512
model-00009-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a745036c263bb875d13ed73653be01cc5fff3cc7d4287e57256f9314c382fc42
3
+ size 4997935584
model-00010-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d5755c9ab2a0fa5e3cd82bf36b6f5541399ae3f1271f3ecc91fa1ba99a17f60
3
+ size 4999211600
model-00011-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cee1d65f9ded3975783c91cc351bc14cc6189a9c90165ad60bbca2bd428b92fa
3
+ size 4997936064
model-00012-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26f9f75843dd38d8f113122574d7feb82ac10fb6e83458a515c7db80c5e8d9be
3
+ size 4999212520
model-00013-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5b4a3c01e8074f0eefb82b7067d7f8dbc02ba284f8074b27cc1654342669af5
3
+ size 4997936232
model-00014-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:80678a88ba9ae9581d34d0f023fb5b56771517676f2a13cf11c7d7b076514465
3
+ size 4999212600
model-00015-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f7d122d91a6cb6caff3e627609a9a3290d5e72bf13e4e5c2333e469ba353cbf
3
+ size 4997936144
model-00016-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b31884c6b7e6de7978253df6166f7669ffd483fad0481a2ab8b5774398a10b18
3
+ size 4979551568
model-00017-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bff26398125fcc8362618371e829fa295f07d1cfaf7e618e435909c31e3c2107
3
+ size 4994003616
model-00018-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb1a51fa9d8928cdbb2c90877ba4600073411373d35c7b085a77f356f91e3fce
3
+ size 4995314936
model-00019-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f543e5b87aefb2948a709a01eda7cb65c57aa97ac8ef3e3fa59e7caa1b85e6f9
3
+ size 4999212224
model-00020-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b5a5db1572dbfc92f923bdf1fe6382d17e5842cc23ac3d3f3cde1a8fcf17453
3
+ size 4997936448
model-00021-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:252d08ee894be03d88c91a92331439dbcf214faa281814337f26a3950a36f5fb
3
+ size 4999212312
model-00022-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8559079efdf411202af8c9e064fc371f1f344dcafa6d0583c47a3a647a277538
3
+ size 4997936408
model-00023-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d33662eb868a088e10e7bb93743519f38651f23a18d337481c9532d67114f73b
3
+ size 4999212344
model-00024-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb43249023af0d8e2ca3d8686ffe6abacfe018854e24d9539f4ce3a11ffe3a0e
3
+ size 4997936408
model-00025-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33c6ddeec94684fad434365c2d15ea2ddd3ba8d05752ebf8621eed552a99fd49
3
+ size 4999212344
model-00026-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3f46b966bbbbf2175ac7e8ea44aa97f7e631f540866c779106253a8d0daa3888
3
+ size 4997936408
model-00027-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:955c49e85e5f11b372fc53cf88847e84ac0ea012212ac8670652ab7323136fba
3
+ size 4999212352
model-00028-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fddfb9fcdd8269d13a0341acf316129918b962c9c3e8dd4ce08b3660830d185
3
+ size 4997936408
model-00029-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:272ff0f904358ca28bdb5e4ce0347eeb42e598cbc27cdffb88c81090d16bf8be
3
+ size 4999212384
model-00030-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca459b499989bcc59a7ad6d5181dfbabe8ca0677df24c1ebaf7451599ecc25f0
3
+ size 4997936360
model-00031-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:20d1fd81df66826e6a46b6bbfb3b7b00167557834d2f595e888ab54d9da8dfe1
3
+ size 4999212472
model-00032-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14ddf0cd19f3fbe604551bbb811e6f2e471b7f2e54d893139ab74e405a9e235b
3
+ size 4997936280
model-00033-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:760edc240ba52619cee1aa6a34ca541356f67040b8154dcdb767fafe2bbcb9b7
3
+ size 4999212552
model-00034-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fd8130a58a071a8cdb905ab333ea9d5fb85ebaab86dcc58af6c187a437894fb
3
+ size 4997936200
model-00035-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42dd2592c6ef05d795dd435065ad3495c695f77c5c01257e35e80ede11116a8f
3
+ size 4999212624
model-00036-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:004f20ab8978340710734de96b86b654c1be89a08d15be550752877b7040164a
3
+ size 4997936120
model-00037-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5224716c2e53a6d19d4a4f7560eac3a2c67311843e97ee88fb5cf1ab3b4c24d
3
+ size 4958598872
model-00038-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:00268a85ad858c26caa5f2c4ac54228d8a5bdfedba772b1743ebc199933a778e
3
+ size 4991362816
model-00039-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:716006d01133f2c80a7254ef5643414a9e99ccf9370ee024b4d07ce031de5f5a
3
+ size 4995314912
model-00040-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:51d5e5a4b166bdf70ae7c9139fa73df3e69f116c0997a97708329660666c5e85
3
+ size 4999212248
model-00041-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d536fbd139804453d1423eff81a86973490e9f2de4f37a574f8b34f985cf20c
3
+ size 4997936424
model-00042-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b282381cb9a9776c5cea197f5a6207ef6e0430e2e508101995a6fabca2c4f13e
3
+ size 4999212336
model-00043-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8edc3e5420fa4e80c167f4962087c6d28863465da025583495df3392ac6898e
3
+ size 4997936408
model-00044-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a4c17f2241e8333747d1ccaf1f89d397c394e1b4ce3884205203b0f9be166fc
3
+ size 4999212344
model-00045-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5594a57bddae13838243aef1db57d75a5c839368f411467833aae014a2e6297e
3
+ size 4997936408
model-00046-of-00047.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87f8720e439c3a912a60495f3d6d9c3affa4147a00966b9424a2fc0dfe41f4bb
3
+ size 4999212344