Hugging Face's logo

aisi-whitebox-red-team
/
davor

Safetensors
nemotron-nas
custom_code
Model card Files
xet
Community
sdtblck commited on
Commit
15cc2aa
·
verified ·
1 Parent(s): 31dc922

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. block_config.py +118 -0
  2. chat_template.jinja +46 -0
  3. config.json +1484 -0
  4. configuration_decilm.py +65 -0
  5. generation_config.json +11 -0
  6. model-00001-of-00044.safetensors +3 -0
  7. model-00002-of-00044.safetensors +3 -0
  8. model-00003-of-00044.safetensors +3 -0
  9. model-00004-of-00044.safetensors +3 -0
  10. model-00005-of-00044.safetensors +3 -0
  11. model-00006-of-00044.safetensors +3 -0
  12. model-00007-of-00044.safetensors +3 -0
  13. model-00008-of-00044.safetensors +3 -0
  14. model-00009-of-00044.safetensors +3 -0
  15. model-00010-of-00044.safetensors +3 -0
  16. model-00011-of-00044.safetensors +3 -0
  17. model-00012-of-00044.safetensors +3 -0
  18. model-00013-of-00044.safetensors +3 -0
  19. model-00014-of-00044.safetensors +3 -0
  20. model-00015-of-00044.safetensors +3 -0
  21. model-00016-of-00044.safetensors +3 -0
  22. model-00017-of-00044.safetensors +3 -0
  23. model-00018-of-00044.safetensors +3 -0
  24. model-00019-of-00044.safetensors +3 -0
  25. model-00020-of-00044.safetensors +3 -0
  26. model-00021-of-00044.safetensors +3 -0
  27. model-00022-of-00044.safetensors +3 -0
  28. model-00023-of-00044.safetensors +3 -0
  29. model-00024-of-00044.safetensors +3 -0
  30. model-00025-of-00044.safetensors +3 -0
  31. model-00026-of-00044.safetensors +3 -0
  32. model-00027-of-00044.safetensors +3 -0
  33. model-00028-of-00044.safetensors +3 -0
  34. model-00029-of-00044.safetensors +3 -0
  35. model-00030-of-00044.safetensors +3 -0
  36. model-00031-of-00044.safetensors +3 -0
  37. model-00032-of-00044.safetensors +3 -0
  38. model-00033-of-00044.safetensors +3 -0
  39. model-00034-of-00044.safetensors +3 -0
  40. model-00035-of-00044.safetensors +3 -0
  41. model-00036-of-00044.safetensors +3 -0
  42. model-00037-of-00044.safetensors +3 -0
  43. model-00038-of-00044.safetensors +3 -0
  44. model-00039-of-00044.safetensors +3 -0
  45. model-00040-of-00044.safetensors +3 -0
  46. model-00041-of-00044.safetensors +3 -0
  47. model-00042-of-00044.safetensors +3 -0
  48. model-00043-of-00044.safetensors +3 -0
  49. model-00044-of-00044.safetensors +3 -0
  50. model.safetensors.index.json +576 -0
block_config.py ADDED
@@ -0,0 +1,118 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import dataclasses
2
+ import json
3
+ import warnings
4
+ from dataclasses import dataclass, MISSING
5
+ from functools import partial
6
+ from typing import Optional, Any
7
+
8
+
9
+ @partial(dataclass, frozen=True, kw_only=True)
10
+ class JsonComparable:
11
+ def to_json(self) -> str:
12
+ return json.dumps(dataclasses.asdict(self))
13
+
14
+ def __eq__(self, other: "JsonComparable") -> bool:
15
+ return self.to_json() == other.to_json()
16
+
17
+ def __hash__(self) -> int:
18
+ return hash(self.to_json())
19
+
20
+ def __lt__(self, other: "JsonComparable") -> bool:
21
+ return self.to_json() < other.to_json()
22
+
23
+
24
+ @partial(dataclass, frozen=True, kw_only=True)
25
+ class SubblockConfig(JsonComparable):
26
+ no_op: bool = False
27
+ replace_with_linear: bool = False
28
+ sparsify: Optional[list[str]] = None
29
+
30
+ def __post_init__(self):
31
+ assert not (self.no_op and self.replace_with_linear)
32
+
33
+ def _force_setattr(self, name: str, value: Any) -> None:
34
+ """
35
+ Set an attribute even in frozen dataclasses.
36
+ Use only inside __post_init__!
37
+ """
38
+ object.__setattr__(self, name, value)
39
+
40
+
41
+ @partial(dataclass, frozen=True, kw_only=True)
42
+ class AttentionConfig(SubblockConfig):
43
+ n_heads_in_group: Optional[int] = None
44
+ window_length: Optional[int] = None
45
+ num_sink_tokens: Optional[int] = None
46
+ use_prefill_window_in_sink_attention: bool = False
47
+ unshifted_sink: bool = False
48
+
49
+ def __post_init__(self):
50
+ super().__post_init__()
51
+ assert not (self.no_op and self.replace_with_linear)
52
+
53
+ if self.no_op or self.replace_with_linear:
54
+ for irrelevant_att in ["n_heads_in_group", "window_length", "num_sink_tokens"]:
55
+ self._force_setattr(irrelevant_att, None)
56
+ else:
57
+ assert self.n_heads_in_group is not None
58
+
59
+ if self.is_sink:
60
+ assert not (self.unshifted_sink and self.use_prefill_window_in_sink_attention), \
61
+ ("Unshifted sink uses its own kind of explicit masking, not standard window. "
62
+ "Set use_prefill_window_in_sink_attention to False.")
63
+ assert not (self.num_sink_tokens == 0 and not self.unshifted_sink), \
64
+ "Fake sink attention with 0 sink tokens is only supported with unshifted_sink=True"
65
+
66
+ @property
67
+ def prefill_sliding_window(self) -> Optional[int]:
68
+ if self.window_length is not None:
69
+ if not self.is_sink or self.use_prefill_window_in_sink_attention:
70
+ return self.window_length
71
+ return None
72
+
73
+ @property
74
+ def is_sliding(self) -> bool:
75
+ return self.prefill_sliding_window is not None
76
+
77
+ @property
78
+ def is_sink(self) -> bool:
79
+ return (
80
+ (self.window_length is not None)
81
+ and
82
+ (self.num_sink_tokens is not None)
83
+ )
84
+
85
+
86
+ @partial(dataclass, frozen=True, kw_only=True)
87
+ class FFNConfig(SubblockConfig):
88
+ ffn_mult: Optional[float] = None
89
+
90
+ def __post_init__(self):
91
+ super().__post_init__()
92
+ if self.no_op or self.replace_with_linear:
93
+ self._force_setattr("ffn_mult", None)
94
+ else:
95
+ assert self.ffn_mult is not None
96
+ self._force_setattr("ffn_mult", round(self.ffn_mult, 6))
97
+
98
+
99
+ @partial(dataclass, frozen=True, kw_only=True)
100
+ class BlockConfig(JsonComparable):
101
+ attention: AttentionConfig = MISSING
102
+ ffn: FFNConfig = MISSING
103
+
104
+ def __post_init__(self):
105
+ """
106
+ Init subblock dataclasses from dicts
107
+ """
108
+ for subblock_name in dataclasses.fields(self):
109
+ subblock_config = getattr(self, subblock_name.name)
110
+ if isinstance(subblock_config, dict):
111
+ subblock_fields = [field.name for field in dataclasses.fields(subblock_name.type)]
112
+ unsupported_fields = [field_name for field_name in subblock_config.keys()
113
+ if field_name not in subblock_fields]
114
+ if len(unsupported_fields) > 0:
115
+ warnings.warn(f"Removed unsupported fields {unsupported_fields} from {subblock_name.type.__name__}")
116
+ subblock_config = {k: v for k, v in subblock_config.items() if k not in unsupported_fields}
117
+ object.__setattr__(self, subblock_name.name,
118
+ subblock_name.type(**subblock_config)) # __setattr__ to overcome frozen=True
chat_template.jinja ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if messages[0]["role"] == "system" -%}
2
+ {%- set system_message = messages[0]["content"] | trim -%}
3
+ {%- set messages = messages[1:] -%}
4
+ {%- else -%}
5
+ {%- set system_message = "" -%}
6
+ {%- endif -%}
7
+ {%- if tools is not none -%}
8
+ {{- "<|begin_of_text|><|start_header_id|>system<|end_header_id|>" + "\n\n" + system_message -}}
9
+ {{- "\n\n" if system_message else "" -}}
10
+ {{- "<AVAILABLE_TOOLS>[" -}}
11
+ {%- for t in tools -%}
12
+ {{- (t.function if t.function is defined else t) | tojson() -}}
13
+ {{- ", " if not loop.last else "" -}}
14
+ {%- endfor -%}
15
+ {{- "]</AVAILABLE_TOOLS>" -}}
16
+ {{- "<|eot_id|>" -}}
17
+ {%- else -%}
18
+ {{- "<|begin_of_text|><|start_header_id|>system<|end_header_id|>" + "\n\n" + system_message + "<|eot_id|>" -}}
19
+ {%- endif -%}
20
+ {%- for message in messages -%}
21
+ {%- if message["role"] == "user" -%}
22
+ {{- "<|start_header_id|>user<|end_header_id|>" + "\n\n" + message["content"] | trim + "<|eot_id|>" -}}
23
+ {%- elif message["role"] == "tool" -%}
24
+ {%- set tool_response = "<TOOL_RESPONSE>[" + message["content"] | trim + "]</TOOL_RESPONSE>" -%}
25
+ {{- "<|start_header_id|>user<|end_header_id|>" + "\n\n" + tool_response + "<|eot_id|>" -}}
26
+ {%- elif message["role"] == "assistant" and message.get("tool_calls") is not none -%}
27
+ {%- set tool_calls = message["tool_calls"] -%}
28
+ {{- "<|start_header_id|>assistant<|end_header_id|>" + "\n\n" + "<TOOLCALL>[" -}}
29
+ {%- for tool_call in tool_calls -%}
30
+ {{- "{" + "\"name\": \"" + tool_call.function.name + "\", \"arguments\": " + tool_call.function.arguments | tojson + "}" -}}
31
+ {%- if not loop.last -%}
32
+ {{- ", " -}}
33
+ {%- else -%}
34
+ {{- "]</TOOLCALL>" + "<|eot_id|>" -}}
35
+ {%- endif -%}
36
+ {%- endfor -%}
37
+ {%- elif message["role"] == "assistant" -%}
38
+ {{- "<|start_header_id|>assistant<|end_header_id|>" + "\n\n" -}}
39
+ {%- generation %}
40
+ {{- message["content"] | trim + "<|eot_id|>" -}}
41
+ {%- endgeneration %}
42
+ {%- endif -%}
43
+ {%- endfor -%}
44
+ {%- if add_generation_prompt -%}
45
+ {{- "<|start_header_id|>assistant<|end_header_id|>" + "\n\n" -}}
46
+ {%- endif -%}
config.json ADDED
@@ -0,0 +1,1484 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DeciLMForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_decilm.DeciLMConfig",
9
+ "AutoModelForCausalLM": "modeling_decilm.DeciLMForCausalLM"
10
+ },
11
+ "block_configs": [
12
+ {
13
+ "attention": {
14
+ "n_heads_in_group": 8,
15
+ "no_op": false,
16
+ "num_sink_tokens": null,
17
+ "replace_with_linear": false,
18
+ "sparsify": null,
19
+ "unshifted_sink": false,
20
+ "use_prefill_window_in_sink_attention": false,
21
+ "window_length": null
22
+ },
23
+ "ffn": {
24
+ "ffn_mult": 2.625,
25
+ "no_op": false,
26
+ "replace_with_linear": false,
27
+ "sparsify": null
28
+ }
29
+ },
30
+ {
31
+ "attention": {
32
+ "n_heads_in_group": 8,
33
+ "no_op": false,
34
+ "num_sink_tokens": null,
35
+ "replace_with_linear": false,
36
+ "sparsify": null,
37
+ "unshifted_sink": false,
38
+ "use_prefill_window_in_sink_attention": false,
39
+ "window_length": null
40
+ },
41
+ "ffn": {
42
+ "ffn_mult": 5.25,
43
+ "no_op": false,
44
+ "replace_with_linear": false,
45
+ "sparsify": null
46
+ }
47
+ },
48
+ {
49
+ "attention": {
50
+ "n_heads_in_group": 8,
51
+ "no_op": false,
52
+ "num_sink_tokens": null,
53
+ "replace_with_linear": false,
54
+ "sparsify": null,
55
+ "unshifted_sink": false,
56
+ "use_prefill_window_in_sink_attention": false,
57
+ "window_length": null
58
+ },
59
+ "ffn": {
60
+ "ffn_mult": 5.25,
61
+ "no_op": false,
62
+ "replace_with_linear": false,
63
+ "sparsify": null
64
+ }
65
+ },
66
+ {
67
+ "attention": {
68
+ "n_heads_in_group": 8,
69
+ "no_op": false,
70
+ "num_sink_tokens": null,
71
+ "replace_with_linear": false,
72
+ "sparsify": null,
73
+ "unshifted_sink": false,
74
+ "use_prefill_window_in_sink_attention": false,
75
+ "window_length": null
76
+ },
77
+ "ffn": {
78
+ "ffn_mult": 5.25,
79
+ "no_op": false,
80
+ "replace_with_linear": false,
81
+ "sparsify": null
82
+ }
83
+ },
84
+ {
85
+ "attention": {
86
+ "n_heads_in_group": 8,
87
+ "no_op": false,
88
+ "num_sink_tokens": null,
89
+ "replace_with_linear": false,
90
+ "sparsify": null,
91
+ "unshifted_sink": false,
92
+ "use_prefill_window_in_sink_attention": false,
93
+ "window_length": null
94
+ },
95
+ "ffn": {
96
+ "ffn_mult": 5.25,
97
+ "no_op": false,
98
+ "replace_with_linear": false,
99
+ "sparsify": null
100
+ }
101
+ },
102
+ {
103
+ "attention": {
104
+ "n_heads_in_group": 8,
105
+ "no_op": false,
106
+ "num_sink_tokens": null,
107
+ "replace_with_linear": false,
108
+ "sparsify": null,
109
+ "unshifted_sink": false,
110
+ "use_prefill_window_in_sink_attention": false,
111
+ "window_length": null
112
+ },
113
+ "ffn": {
114
+ "ffn_mult": 5.25,
115
+ "no_op": false,
116
+ "replace_with_linear": false,
117
+ "sparsify": null
118
+ }
119
+ },
120
+ {
121
+ "attention": {
122
+ "n_heads_in_group": null,
123
+ "no_op": true,
124
+ "num_sink_tokens": null,
125
+ "replace_with_linear": false,
126
+ "sparsify": null,
127
+ "unshifted_sink": false,
128
+ "use_prefill_window_in_sink_attention": false,
129
+ "window_length": null
130
+ },
131
+ "ffn": {
132
+ "ffn_mult": 2.625,
133
+ "no_op": false,
134
+ "replace_with_linear": false,
135
+ "sparsify": null
136
+ }
137
+ },
138
+ {
139
+ "attention": {
140
+ "n_heads_in_group": null,
141
+ "no_op": true,
142
+ "num_sink_tokens": null,
143
+ "replace_with_linear": false,
144
+ "sparsify": null,
145
+ "unshifted_sink": false,
146
+ "use_prefill_window_in_sink_attention": false,
147
+ "window_length": null
148
+ },
149
+ "ffn": {
150
+ "ffn_mult": 2.625,
151
+ "no_op": false,
152
+ "replace_with_linear": false,
153
+ "sparsify": null
154
+ }
155
+ },
156
+ {
157
+ "attention": {
158
+ "n_heads_in_group": 8,
159
+ "no_op": false,
160
+ "num_sink_tokens": null,
161
+ "replace_with_linear": false,
162
+ "sparsify": null,
163
+ "unshifted_sink": false,
164
+ "use_prefill_window_in_sink_attention": false,
165
+ "window_length": null
166
+ },
167
+ "ffn": {
168
+ "ffn_mult": 5.25,
169
+ "no_op": false,
170
+ "replace_with_linear": false,
171
+ "sparsify": null
172
+ }
173
+ },
174
+ {
175
+ "attention": {
176
+ "n_heads_in_group": 8,
177
+ "no_op": false,
178
+ "num_sink_tokens": null,
179
+ "replace_with_linear": false,
180
+ "sparsify": null,
181
+ "unshifted_sink": false,
182
+ "use_prefill_window_in_sink_attention": false,
183
+ "window_length": null
184
+ },
185
+ "ffn": {
186
+ "ffn_mult": 5.25,
187
+ "no_op": false,
188
+ "replace_with_linear": false,
189
+ "sparsify": null
190
+ }
191
+ },
192
+ {
193
+ "attention": {
194
+ "n_heads_in_group": 8,
195
+ "no_op": false,
196
+ "num_sink_tokens": null,
197
+ "replace_with_linear": false,
198
+ "sparsify": null,
199
+ "unshifted_sink": false,
200
+ "use_prefill_window_in_sink_attention": false,
201
+ "window_length": null
202
+ },
203
+ "ffn": {
204
+ "ffn_mult": 5.25,
205
+ "no_op": false,
206
+ "replace_with_linear": false,
207
+ "sparsify": null
208
+ }
209
+ },
210
+ {
211
+ "attention": {
212
+ "n_heads_in_group": null,
213
+ "no_op": true,
214
+ "num_sink_tokens": null,
215
+ "replace_with_linear": false,
216
+ "sparsify": null,
217
+ "unshifted_sink": false,
218
+ "use_prefill_window_in_sink_attention": false,
219
+ "window_length": null
220
+ },
221
+ "ffn": {
222
+ "ffn_mult": 3.28125,
223
+ "no_op": false,
224
+ "replace_with_linear": false,
225
+ "sparsify": null
226
+ }
227
+ },
228
+ {
229
+ "attention": {
230
+ "n_heads_in_group": 8,
231
+ "no_op": false,
232
+ "num_sink_tokens": null,
233
+ "replace_with_linear": false,
234
+ "sparsify": null,
235
+ "unshifted_sink": false,
236
+ "use_prefill_window_in_sink_attention": false,
237
+ "window_length": null
238
+ },
239
+ "ffn": {
240
+ "ffn_mult": 5.25,
241
+ "no_op": false,
242
+ "replace_with_linear": false,
243
+ "sparsify": null
244
+ }
245
+ },
246
+ {
247
+ "attention": {
248
+ "n_heads_in_group": 8,
249
+ "no_op": false,
250
+ "num_sink_tokens": null,
251
+ "replace_with_linear": false,
252
+ "sparsify": null,
253
+ "unshifted_sink": false,
254
+ "use_prefill_window_in_sink_attention": false,
255
+ "window_length": null
256
+ },
257
+ "ffn": {
258
+ "ffn_mult": 5.25,
259
+ "no_op": false,
260
+ "replace_with_linear": false,
261
+ "sparsify": null
262
+ }
263
+ },
264
+ {
265
+ "attention": {
266
+ "n_heads_in_group": 8,
267
+ "no_op": false,
268
+ "num_sink_tokens": null,
269
+ "replace_with_linear": false,
270
+ "sparsify": null,
271
+ "unshifted_sink": false,
272
+ "use_prefill_window_in_sink_attention": false,
273
+ "window_length": null
274
+ },
275
+ "ffn": {
276
+ "ffn_mult": 5.25,
277
+ "no_op": false,
278
+ "replace_with_linear": false,
279
+ "sparsify": null
280
+ }
281
+ },
282
+ {
283
+ "attention": {
284
+ "n_heads_in_group": 8,
285
+ "no_op": false,
286
+ "num_sink_tokens": null,
287
+ "replace_with_linear": false,
288
+ "sparsify": null,
289
+ "unshifted_sink": false,
290
+ "use_prefill_window_in_sink_attention": false,
291
+ "window_length": null
292
+ },
293
+ "ffn": {
294
+ "ffn_mult": 5.25,
295
+ "no_op": false,
296
+ "replace_with_linear": false,
297
+ "sparsify": null
298
+ }
299
+ },
300
+ {
301
+ "attention": {
302
+ "n_heads_in_group": 8,
303
+ "no_op": false,
304
+ "num_sink_tokens": null,
305
+ "replace_with_linear": false,
306
+ "sparsify": null,
307
+ "unshifted_sink": false,
308
+ "use_prefill_window_in_sink_attention": false,
309
+ "window_length": null
310
+ },
311
+ "ffn": {
312
+ "ffn_mult": 5.25,
313
+ "no_op": false,
314
+ "replace_with_linear": false,
315
+ "sparsify": null
316
+ }
317
+ },
318
+ {
319
+ "attention": {
320
+ "n_heads_in_group": 8,
321
+ "no_op": false,
322
+ "num_sink_tokens": null,
323
+ "replace_with_linear": false,
324
+ "sparsify": null,
325
+ "unshifted_sink": false,
326
+ "use_prefill_window_in_sink_attention": false,
327
+ "window_length": null
328
+ },
329
+ "ffn": {
330
+ "ffn_mult": 5.25,
331
+ "no_op": false,
332
+ "replace_with_linear": false,
333
+ "sparsify": null
334
+ }
335
+ },
336
+ {
337
+ "attention": {
338
+ "n_heads_in_group": 8,
339
+ "no_op": false,
340
+ "num_sink_tokens": null,
341
+ "replace_with_linear": false,
342
+ "sparsify": null,
343
+ "unshifted_sink": false,
344
+ "use_prefill_window_in_sink_attention": false,
345
+ "window_length": null
346
+ },
347
+ "ffn": {
348
+ "ffn_mult": 5.25,
349
+ "no_op": false,
350
+ "replace_with_linear": false,
351
+ "sparsify": null
352
+ }
353
+ },
354
+ {
355
+ "attention": {
356
+ "n_heads_in_group": 8,
357
+ "no_op": false,
358
+ "num_sink_tokens": null,
359
+ "replace_with_linear": false,
360
+ "sparsify": null,
361
+ "unshifted_sink": false,
362
+ "use_prefill_window_in_sink_attention": false,
363
+ "window_length": null
364
+ },
365
+ "ffn": {
366
+ "ffn_mult": 5.25,
367
+ "no_op": false,
368
+ "replace_with_linear": false,
369
+ "sparsify": null
370
+ }
371
+ },
372
+ {
373
+ "attention": {
374
+ "n_heads_in_group": 8,
375
+ "no_op": false,
376
+ "num_sink_tokens": null,
377
+ "replace_with_linear": false,
378
+ "sparsify": null,
379
+ "unshifted_sink": false,
380
+ "use_prefill_window_in_sink_attention": false,
381
+ "window_length": null
382
+ },
383
+ "ffn": {
384
+ "ffn_mult": 5.25,
385
+ "no_op": false,
386
+ "replace_with_linear": false,
387
+ "sparsify": null
388
+ }
389
+ },
390
+ {
391
+ "attention": {
392
+ "n_heads_in_group": 8,
393
+ "no_op": false,
394
+ "num_sink_tokens": null,
395
+ "replace_with_linear": false,
396
+ "sparsify": null,
397
+ "unshifted_sink": false,
398
+ "use_prefill_window_in_sink_attention": false,
399
+ "window_length": null
400
+ },
401
+ "ffn": {
402
+ "ffn_mult": 5.25,
403
+ "no_op": false,
404
+ "replace_with_linear": false,
405
+ "sparsify": null
406
+ }
407
+ },
408
+ {
409
+ "attention": {
410
+ "n_heads_in_group": 8,
411
+ "no_op": false,
412
+ "num_sink_tokens": null,
413
+ "replace_with_linear": false,
414
+ "sparsify": null,
415
+ "unshifted_sink": false,
416
+ "use_prefill_window_in_sink_attention": false,
417
+ "window_length": null
418
+ },
419
+ "ffn": {
420
+ "ffn_mult": 5.25,
421
+ "no_op": false,
422
+ "replace_with_linear": false,
423
+ "sparsify": null
424
+ }
425
+ },
426
+ {
427
+ "attention": {
428
+ "n_heads_in_group": 8,
429
+ "no_op": false,
430
+ "num_sink_tokens": null,
431
+ "replace_with_linear": false,
432
+ "sparsify": null,
433
+ "unshifted_sink": false,
434
+ "use_prefill_window_in_sink_attention": false,
435
+ "window_length": null
436
+ },
437
+ "ffn": {
438
+ "ffn_mult": 5.25,
439
+ "no_op": false,
440
+ "replace_with_linear": false,
441
+ "sparsify": null
442
+ }
443
+ },
444
+ {
445
+ "attention": {
446
+ "n_heads_in_group": 8,
447
+ "no_op": false,
448
+ "num_sink_tokens": null,
449
+ "replace_with_linear": false,
450
+ "sparsify": null,
451
+ "unshifted_sink": false,
452
+ "use_prefill_window_in_sink_attention": false,
453
+ "window_length": null
454
+ },
455
+ "ffn": {
456
+ "ffn_mult": 5.25,
457
+ "no_op": false,
458
+ "replace_with_linear": false,
459
+ "sparsify": null
460
+ }
461
+ },
462
+ {
463
+ "attention": {
464
+ "n_heads_in_group": 8,
465
+ "no_op": false,
466
+ "num_sink_tokens": null,
467
+ "replace_with_linear": false,
468
+ "sparsify": null,
469
+ "unshifted_sink": false,
470
+ "use_prefill_window_in_sink_attention": false,
471
+ "window_length": null
472
+ },
473
+ "ffn": {
474
+ "ffn_mult": 5.25,
475
+ "no_op": false,
476
+ "replace_with_linear": false,
477
+ "sparsify": null
478
+ }
479
+ },
480
+ {
481
+ "attention": {
482
+ "n_heads_in_group": 8,
483
+ "no_op": false,
484
+ "num_sink_tokens": null,
485
+ "replace_with_linear": false,
486
+ "sparsify": null,
487
+ "unshifted_sink": false,
488
+ "use_prefill_window_in_sink_attention": false,
489
+ "window_length": null
490
+ },
491
+ "ffn": {
492
+ "ffn_mult": 5.25,
493
+ "no_op": false,
494
+ "replace_with_linear": false,
495
+ "sparsify": null
496
+ }
497
+ },
498
+ {
499
+ "attention": {
500
+ "n_heads_in_group": 8,
501
+ "no_op": false,
502
+ "num_sink_tokens": null,
503
+ "replace_with_linear": false,
504
+ "sparsify": null,
505
+ "unshifted_sink": false,
506
+ "use_prefill_window_in_sink_attention": false,
507
+ "window_length": null
508
+ },
509
+ "ffn": {
510
+ "ffn_mult": 5.25,
511
+ "no_op": false,
512
+ "replace_with_linear": false,
513
+ "sparsify": null
514
+ }
515
+ },
516
+ {
517
+ "attention": {
518
+ "n_heads_in_group": 8,
519
+ "no_op": false,
520
+ "num_sink_tokens": null,
521
+ "replace_with_linear": false,
522
+ "sparsify": null,
523
+ "unshifted_sink": false,
524
+ "use_prefill_window_in_sink_attention": false,
525
+ "window_length": null
526
+ },
527
+ "ffn": {
528
+ "ffn_mult": 5.25,
529
+ "no_op": false,
530
+ "replace_with_linear": false,
531
+ "sparsify": null
532
+ }
533
+ },
534
+ {
535
+ "attention": {
536
+ "n_heads_in_group": 8,
537
+ "no_op": false,
538
+ "num_sink_tokens": null,
539
+ "replace_with_linear": false,
540
+ "sparsify": null,
541
+ "unshifted_sink": false,
542
+ "use_prefill_window_in_sink_attention": false,
543
+ "window_length": null
544
+ },
545
+ "ffn": {
546
+ "ffn_mult": 5.25,
547
+ "no_op": false,
548
+ "replace_with_linear": false,
549
+ "sparsify": null
550
+ }
551
+ },
552
+ {
553
+ "attention": {
554
+ "n_heads_in_group": 8,
555
+ "no_op": false,
556
+ "num_sink_tokens": null,
557
+ "replace_with_linear": false,
558
+ "sparsify": null,
559
+ "unshifted_sink": false,
560
+ "use_prefill_window_in_sink_attention": false,
561
+ "window_length": null
562
+ },
563
+ "ffn": {
564
+ "ffn_mult": 5.25,
565
+ "no_op": false,
566
+ "replace_with_linear": false,
567
+ "sparsify": null
568
+ }
569
+ },
570
+ {
571
+ "attention": {
572
+ "n_heads_in_group": 8,
573
+ "no_op": false,
574
+ "num_sink_tokens": null,
575
+ "replace_with_linear": false,
576
+ "sparsify": null,
577
+ "unshifted_sink": false,
578
+ "use_prefill_window_in_sink_attention": false,
579
+ "window_length": null
580
+ },
581
+ "ffn": {
582
+ "ffn_mult": 5.25,
583
+ "no_op": false,
584
+ "replace_with_linear": false,
585
+ "sparsify": null
586
+ }
587
+ },
588
+ {
589
+ "attention": {
590
+ "n_heads_in_group": 8,
591
+ "no_op": false,
592
+ "num_sink_tokens": null,
593
+ "replace_with_linear": false,
594
+ "sparsify": null,
595
+ "unshifted_sink": false,
596
+ "use_prefill_window_in_sink_attention": false,
597
+ "window_length": null
598
+ },
599
+ "ffn": {
600
+ "ffn_mult": 5.25,
601
+ "no_op": false,
602
+ "replace_with_linear": false,
603
+ "sparsify": null
604
+ }
605
+ },
606
+ {
607
+ "attention": {
608
+ "n_heads_in_group": 8,
609
+ "no_op": false,
610
+ "num_sink_tokens": null,
611
+ "replace_with_linear": false,
612
+ "sparsify": null,
613
+ "unshifted_sink": false,
614
+ "use_prefill_window_in_sink_attention": false,
615
+ "window_length": null
616
+ },
617
+ "ffn": {
618
+ "ffn_mult": 5.25,
619
+ "no_op": false,
620
+ "replace_with_linear": false,
621
+ "sparsify": null
622
+ }
623
+ },
624
+ {
625
+ "attention": {
626
+ "n_heads_in_group": 8,
627
+ "no_op": false,
628
+ "num_sink_tokens": null,
629
+ "replace_with_linear": false,
630
+ "sparsify": null,
631
+ "unshifted_sink": false,
632
+ "use_prefill_window_in_sink_attention": false,
633
+ "window_length": null
634
+ },
635
+ "ffn": {
636
+ "ffn_mult": 5.25,
637
+ "no_op": false,
638
+ "replace_with_linear": false,
639
+ "sparsify": null
640
+ }
641
+ },
642
+ {
643
+ "attention": {
644
+ "n_heads_in_group": 8,
645
+ "no_op": false,
646
+ "num_sink_tokens": null,
647
+ "replace_with_linear": false,
648
+ "sparsify": null,
649
+ "unshifted_sink": false,
650
+ "use_prefill_window_in_sink_attention": false,
651
+ "window_length": null
652
+ },
653
+ "ffn": {
654
+ "ffn_mult": 5.25,
655
+ "no_op": false,
656
+ "replace_with_linear": false,
657
+ "sparsify": null
658
+ }
659
+ },
660
+ {
661
+ "attention": {
662
+ "n_heads_in_group": 8,
663
+ "no_op": false,
664
+ "num_sink_tokens": null,
665
+ "replace_with_linear": false,
666
+ "sparsify": null,
667
+ "unshifted_sink": false,
668
+ "use_prefill_window_in_sink_attention": false,
669
+ "window_length": null
670
+ },
671
+ "ffn": {
672
+ "ffn_mult": 5.25,
673
+ "no_op": false,
674
+ "replace_with_linear": false,
675
+ "sparsify": null
676
+ }
677
+ },
678
+ {
679
+ "attention": {
680
+ "n_heads_in_group": 8,
681
+ "no_op": false,
682
+ "num_sink_tokens": null,
683
+ "replace_with_linear": false,
684
+ "sparsify": null,
685
+ "unshifted_sink": false,
686
+ "use_prefill_window_in_sink_attention": false,
687
+ "window_length": null
688
+ },
689
+ "ffn": {
690
+ "ffn_mult": 5.25,
691
+ "no_op": false,
692
+ "replace_with_linear": false,
693
+ "sparsify": null
694
+ }
695
+ },
696
+ {
697
+ "attention": {
698
+ "n_heads_in_group": 8,
699
+ "no_op": false,
700
+ "num_sink_tokens": null,
701
+ "replace_with_linear": false,
702
+ "sparsify": null,
703
+ "unshifted_sink": false,
704
+ "use_prefill_window_in_sink_attention": false,
705
+ "window_length": null
706
+ },
707
+ "ffn": {
708
+ "ffn_mult": 5.25,
709
+ "no_op": false,
710
+ "replace_with_linear": false,
711
+ "sparsify": null
712
+ }
713
+ },
714
+ {
715
+ "attention": {
716
+ "n_heads_in_group": 8,
717
+ "no_op": false,
718
+ "num_sink_tokens": null,
719
+ "replace_with_linear": false,
720
+ "sparsify": null,
721
+ "unshifted_sink": false,
722
+ "use_prefill_window_in_sink_attention": false,
723
+ "window_length": null
724
+ },
725
+ "ffn": {
726
+ "ffn_mult": 5.25,
727
+ "no_op": false,
728
+ "replace_with_linear": false,
729
+ "sparsify": null
730
+ }
731
+ },
732
+ {
733
+ "attention": {
734
+ "n_heads_in_group": 8,
735
+ "no_op": false,
736
+ "num_sink_tokens": null,
737
+ "replace_with_linear": false,
738
+ "sparsify": null,
739
+ "unshifted_sink": false,
740
+ "use_prefill_window_in_sink_attention": false,
741
+ "window_length": null
742
+ },
743
+ "ffn": {
744
+ "ffn_mult": 5.25,
745
+ "no_op": false,
746
+ "replace_with_linear": false,
747
+ "sparsify": null
748
+ }
749
+ },
750
+ {
751
+ "attention": {
752
+ "n_heads_in_group": 8,
753
+ "no_op": false,
754
+ "num_sink_tokens": null,
755
+ "replace_with_linear": false,
756
+ "sparsify": null,
757
+ "unshifted_sink": false,
758
+ "use_prefill_window_in_sink_attention": false,
759
+ "window_length": null
760
+ },
761
+ "ffn": {
762
+ "ffn_mult": 5.25,
763
+ "no_op": false,
764
+ "replace_with_linear": false,
765
+ "sparsify": null
766
+ }
767
+ },
768
+ {
769
+ "attention": {
770
+ "n_heads_in_group": null,
771
+ "no_op": true,
772
+ "num_sink_tokens": null,
773
+ "replace_with_linear": false,
774
+ "sparsify": null,
775
+ "unshifted_sink": false,
776
+ "use_prefill_window_in_sink_attention": false,
777
+ "window_length": null
778
+ },
779
+ "ffn": {
780
+ "ffn_mult": 1.3125,
781
+ "no_op": false,
782
+ "replace_with_linear": false,
783
+ "sparsify": null
784
+ }
785
+ },
786
+ {
787
+ "attention": {
788
+ "n_heads_in_group": null,
789
+ "no_op": true,
790
+ "num_sink_tokens": null,
791
+ "replace_with_linear": false,
792
+ "sparsify": null,
793
+ "unshifted_sink": false,
794
+ "use_prefill_window_in_sink_attention": false,
795
+ "window_length": null
796
+ },
797
+ "ffn": {
798
+ "ffn_mult": 2.625,
799
+ "no_op": false,
800
+ "replace_with_linear": false,
801
+ "sparsify": null
802
+ }
803
+ },
804
+ {
805
+ "attention": {
806
+ "n_heads_in_group": null,
807
+ "no_op": true,
808
+ "num_sink_tokens": null,
809
+ "replace_with_linear": false,
810
+ "sparsify": null,
811
+ "unshifted_sink": false,
812
+ "use_prefill_window_in_sink_attention": false,
813
+ "window_length": null
814
+ },
815
+ "ffn": {
816
+ "ffn_mult": 2.625,
817
+ "no_op": false,
818
+ "replace_with_linear": false,
819
+ "sparsify": null
820
+ }
821
+ },
822
+ {
823
+ "attention": {
824
+ "n_heads_in_group": null,
825
+ "no_op": true,
826
+ "num_sink_tokens": null,
827
+ "replace_with_linear": false,
828
+ "sparsify": null,
829
+ "unshifted_sink": false,
830
+ "use_prefill_window_in_sink_attention": false,
831
+ "window_length": null
832
+ },
833
+ "ffn": {
834
+ "ffn_mult": 1.3125,
835
+ "no_op": false,
836
+ "replace_with_linear": false,
837
+ "sparsify": null
838
+ }
839
+ },
840
+ {
841
+ "attention": {
842
+ "n_heads_in_group": null,
843
+ "no_op": true,
844
+ "num_sink_tokens": null,
845
+ "replace_with_linear": false,
846
+ "sparsify": null,
847
+ "unshifted_sink": false,
848
+ "use_prefill_window_in_sink_attention": false,
849
+ "window_length": null
850
+ },
851
+ "ffn": {
852
+ "ffn_mult": 5.25,
853
+ "no_op": false,
854
+ "replace_with_linear": false,
855
+ "sparsify": null
856
+ }
857
+ },
858
+ {
859
+ "attention": {
860
+ "n_heads_in_group": null,
861
+ "no_op": true,
862
+ "num_sink_tokens": null,
863
+ "replace_with_linear": false,
864
+ "sparsify": null,
865
+ "unshifted_sink": false,
866
+ "use_prefill_window_in_sink_attention": false,
867
+ "window_length": null
868
+ },
869
+ "ffn": {
870
+ "ffn_mult": 1.3125,
871
+ "no_op": false,
872
+ "replace_with_linear": false,
873
+ "sparsify": null
874
+ }
875
+ },
876
+ {
877
+ "attention": {
878
+ "n_heads_in_group": null,
879
+ "no_op": true,
880
+ "num_sink_tokens": null,
881
+ "replace_with_linear": false,
882
+ "sparsify": null,
883
+ "unshifted_sink": false,
884
+ "use_prefill_window_in_sink_attention": false,
885
+ "window_length": null
886
+ },
887
+ "ffn": {
888
+ "ffn_mult": 2.625,
889
+ "no_op": false,
890
+ "replace_with_linear": false,
891
+ "sparsify": null
892
+ }
893
+ },
894
+ {
895
+ "attention": {
896
+ "n_heads_in_group": null,
897
+ "no_op": true,
898
+ "num_sink_tokens": null,
899
+ "replace_with_linear": false,
900
+ "sparsify": null,
901
+ "unshifted_sink": false,
902
+ "use_prefill_window_in_sink_attention": false,
903
+ "window_length": null
904
+ },
905
+ "ffn": {
906
+ "ffn_mult": 1.3125,
907
+ "no_op": false,
908
+ "replace_with_linear": false,
909
+ "sparsify": null
910
+ }
911
+ },
912
+ {
913
+ "attention": {
914
+ "n_heads_in_group": null,
915
+ "no_op": true,
916
+ "num_sink_tokens": null,
917
+ "replace_with_linear": false,
918
+ "sparsify": null,
919
+ "unshifted_sink": false,
920
+ "use_prefill_window_in_sink_attention": false,
921
+ "window_length": null
922
+ },
923
+ "ffn": {
924
+ "ffn_mult": 1.3125,
925
+ "no_op": false,
926
+ "replace_with_linear": false,
927
+ "sparsify": null
928
+ }
929
+ },
930
+ {
931
+ "attention": {
932
+ "n_heads_in_group": null,
933
+ "no_op": true,
934
+ "num_sink_tokens": null,
935
+ "replace_with_linear": false,
936
+ "sparsify": null,
937
+ "unshifted_sink": false,
938
+ "use_prefill_window_in_sink_attention": false,
939
+ "window_length": null
940
+ },
941
+ "ffn": {
942
+ "ffn_mult": 1.3125,
943
+ "no_op": false,
944
+ "replace_with_linear": false,
945
+ "sparsify": null
946
+ }
947
+ },
948
+ {
949
+ "attention": {
950
+ "n_heads_in_group": 8,
951
+ "no_op": false,
952
+ "num_sink_tokens": null,
953
+ "replace_with_linear": false,
954
+ "sparsify": null,
955
+ "unshifted_sink": false,
956
+ "use_prefill_window_in_sink_attention": false,
957
+ "window_length": null
958
+ },
959
+ "ffn": {
960
+ "ffn_mult": 5.25,
961
+ "no_op": false,
962
+ "replace_with_linear": false,
963
+ "sparsify": null
964
+ }
965
+ },
966
+ {
967
+ "attention": {
968
+ "n_heads_in_group": null,
969
+ "no_op": true,
970
+ "num_sink_tokens": null,
971
+ "replace_with_linear": false,
972
+ "sparsify": null,
973
+ "unshifted_sink": false,
974
+ "use_prefill_window_in_sink_attention": false,
975
+ "window_length": null
976
+ },
977
+ "ffn": {
978
+ "ffn_mult": 1.3125,
979
+ "no_op": false,
980
+ "replace_with_linear": false,
981
+ "sparsify": null
982
+ }
983
+ },
984
+ {
985
+ "attention": {
986
+ "n_heads_in_group": null,
987
+ "no_op": true,
988
+ "num_sink_tokens": null,
989
+ "replace_with_linear": false,
990
+ "sparsify": null,
991
+ "unshifted_sink": false,
992
+ "use_prefill_window_in_sink_attention": false,
993
+ "window_length": null
994
+ },
995
+ "ffn": {
996
+ "ffn_mult": 1.0,
997
+ "no_op": false,
998
+ "replace_with_linear": false,
999
+ "sparsify": null
1000
+ }
1001
+ },
1002
+ {
1003
+ "attention": {
1004
+ "n_heads_in_group": null,
1005
+ "no_op": true,
1006
+ "num_sink_tokens": null,
1007
+ "replace_with_linear": false,
1008
+ "sparsify": null,
1009
+ "unshifted_sink": false,
1010
+ "use_prefill_window_in_sink_attention": false,
1011
+ "window_length": null
1012
+ },
1013
+ "ffn": {
1014
+ "ffn_mult": 1.0,
1015
+ "no_op": false,
1016
+ "replace_with_linear": false,
1017
+ "sparsify": null
1018
+ }
1019
+ },
1020
+ {
1021
+ "attention": {
1022
+ "n_heads_in_group": null,
1023
+ "no_op": true,
1024
+ "num_sink_tokens": null,
1025
+ "replace_with_linear": false,
1026
+ "sparsify": null,
1027
+ "unshifted_sink": false,
1028
+ "use_prefill_window_in_sink_attention": false,
1029
+ "window_length": null
1030
+ },
1031
+ "ffn": {
1032
+ "ffn_mult": 1.3125,
1033
+ "no_op": false,
1034
+ "replace_with_linear": false,
1035
+ "sparsify": null
1036
+ }
1037
+ },
1038
+ {
1039
+ "attention": {
1040
+ "n_heads_in_group": null,
1041
+ "no_op": true,
1042
+ "num_sink_tokens": null,
1043
+ "replace_with_linear": false,
1044
+ "sparsify": null,
1045
+ "unshifted_sink": false,
1046
+ "use_prefill_window_in_sink_attention": false,
1047
+ "window_length": null
1048
+ },
1049
+ "ffn": {
1050
+ "ffn_mult": 1.0,
1051
+ "no_op": false,
1052
+ "replace_with_linear": false,
1053
+ "sparsify": null
1054
+ }
1055
+ },
1056
+ {
1057
+ "attention": {
1058
+ "n_heads_in_group": null,
1059
+ "no_op": true,
1060
+ "num_sink_tokens": null,
1061
+ "replace_with_linear": false,
1062
+ "sparsify": null,
1063
+ "unshifted_sink": false,
1064
+ "use_prefill_window_in_sink_attention": false,
1065
+ "window_length": null
1066
+ },
1067
+ "ffn": {
1068
+ "ffn_mult": 1.0,
1069
+ "no_op": false,
1070
+ "replace_with_linear": false,
1071
+ "sparsify": null
1072
+ }
1073
+ },
1074
+ {
1075
+ "attention": {
1076
+ "n_heads_in_group": null,
1077
+ "no_op": true,
1078
+ "num_sink_tokens": null,
1079
+ "replace_with_linear": false,
1080
+ "sparsify": null,
1081
+ "unshifted_sink": false,
1082
+ "use_prefill_window_in_sink_attention": false,
1083
+ "window_length": null
1084
+ },
1085
+ "ffn": {
1086
+ "ffn_mult": 1.0,
1087
+ "no_op": false,
1088
+ "replace_with_linear": false,
1089
+ "sparsify": null
1090
+ }
1091
+ },
1092
+ {
1093
+ "attention": {
1094
+ "n_heads_in_group": null,
1095
+ "no_op": true,
1096
+ "num_sink_tokens": null,
1097
+ "replace_with_linear": false,
1098
+ "sparsify": null,
1099
+ "unshifted_sink": false,
1100
+ "use_prefill_window_in_sink_attention": false,
1101
+ "window_length": null
1102
+ },
1103
+ "ffn": {
1104
+ "ffn_mult": 1.3125,
1105
+ "no_op": false,
1106
+ "replace_with_linear": false,
1107
+ "sparsify": null
1108
+ }
1109
+ },
1110
+ {
1111
+ "attention": {
1112
+ "n_heads_in_group": null,
1113
+ "no_op": true,
1114
+ "num_sink_tokens": null,
1115
+ "replace_with_linear": false,
1116
+ "sparsify": null,
1117
+ "unshifted_sink": false,
1118
+ "use_prefill_window_in_sink_attention": false,
1119
+ "window_length": null
1120
+ },
1121
+ "ffn": {
1122
+ "ffn_mult": 1.3125,
1123
+ "no_op": false,
1124
+ "replace_with_linear": false,
1125
+ "sparsify": null
1126
+ }
1127
+ },
1128
+ {
1129
+ "attention": {
1130
+ "n_heads_in_group": null,
1131
+ "no_op": true,
1132
+ "num_sink_tokens": null,
1133
+ "replace_with_linear": false,
1134
+ "sparsify": null,
1135
+ "unshifted_sink": false,
1136
+ "use_prefill_window_in_sink_attention": false,
1137
+ "window_length": null
1138
+ },
1139
+ "ffn": {
1140
+ "ffn_mult": 0.5,
1141
+ "no_op": false,
1142
+ "replace_with_linear": false,
1143
+ "sparsify": null
1144
+ }
1145
+ },
1146
+ {
1147
+ "attention": {
1148
+ "n_heads_in_group": null,
1149
+ "no_op": true,
1150
+ "num_sink_tokens": null,
1151
+ "replace_with_linear": false,
1152
+ "sparsify": null,
1153
+ "unshifted_sink": false,
1154
+ "use_prefill_window_in_sink_attention": false,
1155
+ "window_length": null
1156
+ },
1157
+ "ffn": {
1158
+ "ffn_mult": 0.5,
1159
+ "no_op": false,
1160
+ "replace_with_linear": false,
1161
+ "sparsify": null
1162
+ }
1163
+ },
1164
+ {
1165
+ "attention": {
1166
+ "n_heads_in_group": null,
1167
+ "no_op": true,
1168
+ "num_sink_tokens": null,
1169
+ "replace_with_linear": false,
1170
+ "sparsify": null,
1171
+ "unshifted_sink": false,
1172
+ "use_prefill_window_in_sink_attention": false,
1173
+ "window_length": null
1174
+ },
1175
+ "ffn": {
1176
+ "ffn_mult": 1.0,
1177
+ "no_op": false,
1178
+ "replace_with_linear": false,
1179
+ "sparsify": null
1180
+ }
1181
+ },
1182
+ {
1183
+ "attention": {
1184
+ "n_heads_in_group": null,
1185
+ "no_op": true,
1186
+ "num_sink_tokens": null,
1187
+ "replace_with_linear": false,
1188
+ "sparsify": null,
1189
+ "unshifted_sink": false,
1190
+ "use_prefill_window_in_sink_attention": false,
1191
+ "window_length": null
1192
+ },
1193
+ "ffn": {
1194
+ "ffn_mult": 1.0,
1195
+ "no_op": false,
1196
+ "replace_with_linear": false,
1197
+ "sparsify": null
1198
+ }
1199
+ },
1200
+ {
1201
+ "attention": {
1202
+ "n_heads_in_group": null,
1203
+ "no_op": true,
1204
+ "num_sink_tokens": null,
1205
+ "replace_with_linear": false,
1206
+ "sparsify": null,
1207
+ "unshifted_sink": false,
1208
+ "use_prefill_window_in_sink_attention": false,
1209
+ "window_length": null
1210
+ },
1211
+ "ffn": {
1212
+ "ffn_mult": 0.5,
1213
+ "no_op": false,
1214
+ "replace_with_linear": false,
1215
+ "sparsify": null
1216
+ }
1217
+ },
1218
+ {
1219
+ "attention": {
1220
+ "n_heads_in_group": null,
1221
+ "no_op": true,
1222
+ "num_sink_tokens": null,
1223
+ "replace_with_linear": false,
1224
+ "sparsify": null,
1225
+ "unshifted_sink": false,
1226
+ "use_prefill_window_in_sink_attention": false,
1227
+ "window_length": null
1228
+ },
1229
+ "ffn": {
1230
+ "ffn_mult": 0.5,
1231
+ "no_op": false,
1232
+ "replace_with_linear": false,
1233
+ "sparsify": null
1234
+ }
1235
+ },
1236
+ {
1237
+ "attention": {
1238
+ "n_heads_in_group": null,
1239
+ "no_op": true,
1240
+ "num_sink_tokens": null,
1241
+ "replace_with_linear": false,
1242
+ "sparsify": null,
1243
+ "unshifted_sink": false,
1244
+ "use_prefill_window_in_sink_attention": false,
1245
+ "window_length": null
1246
+ },
1247
+ "ffn": {
1248
+ "ffn_mult": 1.0,
1249
+ "no_op": false,
1250
+ "replace_with_linear": false,
1251
+ "sparsify": null
1252
+ }
1253
+ },
1254
+ {
1255
+ "attention": {
1256
+ "n_heads_in_group": null,
1257
+ "no_op": true,
1258
+ "num_sink_tokens": null,
1259
+ "replace_with_linear": false,
1260
+ "sparsify": null,
1261
+ "unshifted_sink": false,
1262
+ "use_prefill_window_in_sink_attention": false,
1263
+ "window_length": null
1264
+ },
1265
+ "ffn": {
1266
+ "ffn_mult": 0.5,
1267
+ "no_op": false,
1268
+ "replace_with_linear": false,
1269
+ "sparsify": null
1270
+ }
1271
+ },
1272
+ {
1273
+ "attention": {
1274
+ "n_heads_in_group": null,
1275
+ "no_op": true,
1276
+ "num_sink_tokens": null,
1277
+ "replace_with_linear": false,
1278
+ "sparsify": null,
1279
+ "unshifted_sink": false,
1280
+ "use_prefill_window_in_sink_attention": false,
1281
+ "window_length": null
1282
+ },
1283
+ "ffn": {
1284
+ "ffn_mult": 0.5,
1285
+ "no_op": false,
1286
+ "replace_with_linear": false,
1287
+ "sparsify": null
1288
+ }
1289
+ },
1290
+ {
1291
+ "attention": {
1292
+ "n_heads_in_group": 8,
1293
+ "no_op": false,
1294
+ "num_sink_tokens": null,
1295
+ "replace_with_linear": false,
1296
+ "sparsify": null,
1297
+ "unshifted_sink": false,
1298
+ "use_prefill_window_in_sink_attention": false,
1299
+ "window_length": null
1300
+ },
1301
+ "ffn": {
1302
+ "ffn_mult": 5.25,
1303
+ "no_op": false,
1304
+ "replace_with_linear": false,
1305
+ "sparsify": null
1306
+ }
1307
+ },
1308
+ {
1309
+ "attention": {
1310
+ "n_heads_in_group": 8,
1311
+ "no_op": false,
1312
+ "num_sink_tokens": null,
1313
+ "replace_with_linear": false,
1314
+ "sparsify": null,
1315
+ "unshifted_sink": false,
1316
+ "use_prefill_window_in_sink_attention": false,
1317
+ "window_length": null
1318
+ },
1319
+ "ffn": {
1320
+ "ffn_mult": 5.25,
1321
+ "no_op": false,
1322
+ "replace_with_linear": false,
1323
+ "sparsify": null
1324
+ }
1325
+ },
1326
+ {
1327
+ "attention": {
1328
+ "n_heads_in_group": 8,
1329
+ "no_op": false,
1330
+ "num_sink_tokens": null,
1331
+ "replace_with_linear": false,
1332
+ "sparsify": null,
1333
+ "unshifted_sink": false,
1334
+ "use_prefill_window_in_sink_attention": false,
1335
+ "window_length": null
1336
+ },
1337
+ "ffn": {
1338
+ "ffn_mult": 5.25,
1339
+ "no_op": false,
1340
+ "replace_with_linear": false,
1341
+ "sparsify": null
1342
+ }
1343
+ },
1344
+ {
1345
+ "attention": {
1346
+ "n_heads_in_group": 8,
1347
+ "no_op": false,
1348
+ "num_sink_tokens": null,
1349
+ "replace_with_linear": false,
1350
+ "sparsify": null,
1351
+ "unshifted_sink": false,
1352
+ "use_prefill_window_in_sink_attention": false,
1353
+ "window_length": null
1354
+ },
1355
+ "ffn": {
1356
+ "ffn_mult": 5.25,
1357
+ "no_op": false,
1358
+ "replace_with_linear": false,
1359
+ "sparsify": null
1360
+ }
1361
+ },
1362
+ {
1363
+ "attention": {
1364
+ "n_heads_in_group": 8,
1365
+ "no_op": false,
1366
+ "num_sink_tokens": null,
1367
+ "replace_with_linear": false,
1368
+ "sparsify": null,
1369
+ "unshifted_sink": false,
1370
+ "use_prefill_window_in_sink_attention": false,
1371
+ "window_length": null
1372
+ },
1373
+ "ffn": {
1374
+ "ffn_mult": 5.25,
1375
+ "no_op": false,
1376
+ "replace_with_linear": false,
1377
+ "sparsify": null
1378
+ }
1379
+ },
1380
+ {
1381
+ "attention": {
1382
+ "n_heads_in_group": 8,
1383
+ "no_op": false,
1384
+ "num_sink_tokens": null,
1385
+ "replace_with_linear": false,
1386
+ "sparsify": null,
1387
+ "unshifted_sink": false,
1388
+ "use_prefill_window_in_sink_attention": false,
1389
+ "window_length": null
1390
+ },
1391
+ "ffn": {
1392
+ "ffn_mult": 5.25,
1393
+ "no_op": false,
1394
+ "replace_with_linear": false,
1395
+ "sparsify": null
1396
+ }
1397
+ },
1398
+ {
1399
+ "attention": {
1400
+ "n_heads_in_group": 8,
1401
+ "no_op": false,
1402
+ "num_sink_tokens": null,
1403
+ "replace_with_linear": false,
1404
+ "sparsify": null,
1405
+ "unshifted_sink": false,
1406
+ "use_prefill_window_in_sink_attention": false,
1407
+ "window_length": null
1408
+ },
1409
+ "ffn": {
1410
+ "ffn_mult": 5.25,
1411
+ "no_op": false,
1412
+ "replace_with_linear": false,
1413
+ "sparsify": null
1414
+ }
1415
+ },
1416
+ {
1417
+ "attention": {
1418
+ "n_heads_in_group": 8,
1419
+ "no_op": false,
1420
+ "num_sink_tokens": null,
1421
+ "replace_with_linear": false,
1422
+ "sparsify": null,
1423
+ "unshifted_sink": false,
1424
+ "use_prefill_window_in_sink_attention": false,
1425
+ "window_length": null
1426
+ },
1427
+ "ffn": {
1428
+ "ffn_mult": 5.25,
1429
+ "no_op": false,
1430
+ "replace_with_linear": false,
1431
+ "sparsify": null
1432
+ }
1433
+ },
1434
+ {
1435
+ "attention": {
1436
+ "n_heads_in_group": 8,
1437
+ "no_op": false,
1438
+ "num_sink_tokens": null,
1439
+ "replace_with_linear": false,
1440
+ "sparsify": null,
1441
+ "unshifted_sink": false,
1442
+ "use_prefill_window_in_sink_attention": false,
1443
+ "window_length": null
1444
+ },
1445
+ "ffn": {
1446
+ "ffn_mult": 5.25,
1447
+ "no_op": false,
1448
+ "replace_with_linear": false,
1449
+ "sparsify": null
1450
+ }
1451
+ }
1452
+ ],
1453
+ "bos_token_id": 128000,
1454
+ "eos_token_id": [
1455
+ 128001,
1456
+ 128008,
1457
+ 128009
1458
+ ],
1459
+ "hidden_act": "silu",
1460
+ "hidden_size": 8192,
1461
+ "initializer_range": 0.02,
1462
+ "intermediate_size": null,
1463
+ "max_position_embeddings": 131072,
1464
+ "mlp_bias": false,
1465
+ "model_type": "nemotron-nas",
1466
+ "num_attention_heads": 64,
1467
+ "num_hidden_layers": 80,
1468
+ "num_key_value_heads": null,
1469
+ "pretraining_tp": 1,
1470
+ "rms_norm_eps": 1e-05,
1471
+ "rope_scaling": {
1472
+ "factor": 8.0,
1473
+ "high_freq_factor": 4.0,
1474
+ "low_freq_factor": 1.0,
1475
+ "original_max_position_embeddings": 8192,
1476
+ "rope_type": "llama3"
1477
+ },
1478
+ "rope_theta": 500000.0,
1479
+ "tie_word_embeddings": false,
1480
+ "torch_dtype": "float32",
1481
+ "transformers_version": "4.53.3",
1482
+ "use_cache": true,
1483
+ "vocab_size": 128256
1484
+ }
configuration_decilm.py ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2024 Nvidia Corporation. All rights reserved.
3
+ #
4
+ # Licensed under the Apache License, Version 2.0 (the "License");
5
+ # you may not use this file except in compliance with the License.
6
+ # You may obtain a copy of the License at
7
+ #
8
+ # http://www.apache.org/licenses/LICENSE-2.0
9
+ #
10
+ # Unless required by applicable law or agreed to in writing, software
11
+ # distributed under the License is distributed on an "AS IS" BASIS,
12
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
+ # See the License for the specific language governing permissions and
14
+ # limitations under the License.
15
+
16
+ import dataclasses
17
+ import warnings
18
+ from typing import Dict, Any
19
+
20
+ from transformers.utils import is_flash_attn_2_available
21
+
22
+ from .block_config import BlockConfig
23
+ from .transformers_4_44_2__configuration_llama import LlamaConfig
24
+ from .transformers_4_44_2__modeling_rope_utils import \
25
+ rope_config_validation # fake import to make AutoConfig infer the dependency
26
+
27
+ rope_config_validation # this line is here to make sure that auto-formatting doesn't remove the import
28
+
29
+
30
+ class DeciLMConfig(LlamaConfig):
31
+ model_type = "nemotron-nas"
32
+
33
+ def __init__(
34
+ self,
35
+ block_configs: list[dict] | list[BlockConfig] = None,
36
+ **kwargs,
37
+ ):
38
+ attn_implementation = kwargs.pop("attn_implementation", None)
39
+ if attn_implementation is None and is_flash_attn_2_available():
40
+ attn_implementation = "flash_attention_2"
41
+
42
+ if block_configs is not None:
43
+ if isinstance(block_configs[0], dict):
44
+ block_configs = [BlockConfig(**conf) for conf in block_configs]
45
+
46
+ using_unshifted_sink = any([block_config.attention.unshifted_sink for block_config in block_configs])
47
+ if using_unshifted_sink and attn_implementation != "eager":
48
+ warnings.warn("Forcing attn_implementation='eager' since some attention layers use unshifted sink")
49
+ attn_implementation = "eager"
50
+
51
+ super().__init__(attn_implementation=attn_implementation, **kwargs)
52
+
53
+ self.intermediate_size = None
54
+ self.num_key_value_heads = None
55
+
56
+ if block_configs is not None:
57
+ assert len(block_configs) == self.num_hidden_layers
58
+
59
+ self.block_configs: list[BlockConfig] = block_configs
60
+
61
+ def to_dict(self) -> Dict[str, Any]:
62
+ self_dict = super().to_dict()
63
+ if self.block_configs is not None:
64
+ self_dict["block_configs"] = [dataclasses.asdict(conf) for conf in self.block_configs]
65
+ return self_dict
generation_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 128000,
4
+ "do_sample": true,
5
+ "eos_token_id": [
6
+ 128001,
7
+ 128008,
8
+ 128009
9
+ ],
10
+ "transformers_version": "4.53.3"
11
+ }
model-00001-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:947a6e0a58c53c5718ba92a3341b4faf40f944fd0f303c58934fbbbeea0a8ff8
3
+ size 4806738752
model-00002-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4abdb0931d2537e5b1c926ff4484cae467a2d3178db052d51d1302749a2aac4f
3
+ size 4831938024
model-00003-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:834e9b999f724ccc01a456bd3bdf97143acb29fe7a258d96447543eef1c94644
3
+ size 4966156000
model-00004-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc891b7c6cbd817d6a68c8b8a1f144f346765db9680f632b6d4465aeb34bc1d9
3
+ size 4362142872
model-00005-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:131484a50c23bf397ca9f7ea14924625e44d17c747ae17b2a2d14338fbed9d18
3
+ size 4831937920
model-00006-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf114a1c321c31854d2bcdb4eb2cf2afeac41f62042f75a8f4b1042f6222a5aa
3
+ size 4831938144
model-00007-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d04b4373792d844293f708f6528cf1afe3e8db9b379aad2eaecfe5c3cb7ee5a5
3
+ size 4966188856
model-00008-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a0681a00fea32b2da6edf61a6c5756e81bcfbcc0cf8f9ec064706b61f3c46a6d
3
+ size 4915791128
model-00009-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e7d9bf4b19c4a46e85f7011ca3a74297b105facabdc4270cbfeaaedc32f5a099
3
+ size 4630611344
model-00010-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:760736d6ff3be31f0837a08c77b6dd0b07716e39070a9e53081a562d45bb4cc9
3
+ size 4362142880
model-00011-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5aa674d17abc5c72f877ce0ee4368215bc876d133af2ee92f194bc64927510c0
3
+ size 4966188896
model-00012-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03e00d8591bb14ca4b89b57f6f0b98be1f699bcc7f5cbbfbc7a7589d9f365f34
3
+ size 4362142888
model-00013-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad931dd89f6550bb433caa5bb93eed80c40c2e7b8f51c98302da5f60edb4c5b3
3
+ size 4362142880
model-00014-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a563497caeca286091d956821877c9b3d123f9930779a9f2b1a7be85093e1ec7
3
+ size 4966188896
model-00015-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4856a9d1a7db05f3ffcf05a1be55ebe0ea0b57a21ec0ba191d873de58c6a0bdf
3
+ size 4362142888
model-00016-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02f5d608de8c229fc9dd41a4b3003c69e26a9687d3876077353acc3ec1f0f9dc
3
+ size 4362142880
model-00017-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c69db3a3b324fda4fae4af859544b647a0acf33d9b9c241d4fc9b5c9ae8ed1e
3
+ size 4966188896
model-00018-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:081cf33c3540896d02569a2770d9402765773d92143c2efe1805314153ac9c5d
3
+ size 4362142888
model-00019-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8638eff3c1192fb17a4858fd00c52f27193772045ffc22de670f2497f83e9cd9
3
+ size 4362142880
model-00020-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:adff119059ae002e787cd97d7b0167554a05f87a85881204b36b9ad559f22863
3
+ size 4966188896
model-00021-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d26d727ab454581e43d13a59d4ad8c3a99c6f970aab86b89c61e6f269a476ef
3
+ size 4362142888
model-00022-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74c39c02eb45480c8e0882a3a4b9881af7e5d09d2d3f463c52f2a837c4b44e64
3
+ size 4362142880
model-00023-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04f3865584321ec5621f3f3ea55220d51a0238516d98793c87bd57c6e61041f6
3
+ size 4966188896
model-00024-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdafb4418cc5ffac5f87a46c34063357b77726b7fb7d6ff3995b8c737c0a9961
3
+ size 4362142888
model-00025-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10279b256ccb6e649172e740dab73f9c6c9cdc91750701c66f65ca3ceedb6069
3
+ size 4362142880
model-00026-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f35464d9cfc2af265ea507386ce1b12db668b7771bade9c529be807519a38978
3
+ size 4966188896
model-00027-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63578d387cc9941bc7bff2024f937df2340d53a2151e7e5eb1ea06db44c4193c
3
+ size 4362142888
model-00028-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:88520d5fd37f9014c0d5301193bce090cce3f1d0b122abb354c34284d5b9f1cc
3
+ size 4362142880
model-00029-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ba270319ac7bf916210c48952580277f580367bdce0d194ba2ee933bca3cd2a
3
+ size 4966188896
model-00030-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:467c34b866a722a7c5a3102c6772390edc6e22db117656fb18bc960af6b8a61f
3
+ size 4362142888
model-00031-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:011337d0209849f9773cf07df778d7c7cabc7054c2891f48ad95a1e76478add2
3
+ size 4932601336
model-00032-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:14809feeb5516733cfeab0eb8a2ebf7de1b3f01ef4431f332a6e487f611adef5
3
+ size 4697753200
model-00033-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd4b7dc7f5bdee3dc1727a6b277e47c62b08e5d63572151cecac90328c57712a
3
+ size 4127361432
model-00034-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:39db6314f988891b2bb183fd95b03ab34f280fade1d6bdde6e53746ba709b021
3
+ size 4865525704
model-00035-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4154e1d7710b15bdd69703a60e6a26737ad49e2b12e1edd56da23329c4afda07
3
+ size 4832137288
model-00036-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7f39b81b2f5cb535831b16771597651a6b85731f81d0bc528c8b8ec56604e97
3
+ size 4513303944
model-00037-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa78d7d7ec86ff135f9fdd52be9f4746f42cef54a7f53eb812e984c95f74bfb8
3
+ size 4966188896
model-00038-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c71612a0be36eaacb10051e6e4058b5a59913dc193d25fc07d4f9029e11997fa
3
+ size 4362142888
model-00039-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6199afdda299b805a404341f363e6870a46c425861dbe6ebe2f1e9aa257c5d75
3
+ size 4362142880
model-00040-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15633ace97f6fa29224022af4e8a66549344e0d55c8c24e577026fafb70fbce4
3
+ size 4966188896
model-00041-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:588b759a0eb33485e3b0f8caeae179531dc4cb6026f0ffcd4d05faf51f4c5b5c
3
+ size 4362142888
model-00042-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fae3c91f22829d818bee873f665ebdc028622f36a12d01c0d9bb2c11124dd985
3
+ size 4362142880
model-00043-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a616b8fe07e2461e464b6ab105c866f5ababf6473f324aa8e0f107501f9aae20
3
+ size 939557104
model-00044-of-00044.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a59545fc83f5a4dc160bf2a985233907549d5a67bfa40dac0a8b05c564f34f3
3
+ size 4202692736
model.safetensors.index.json ADDED
@@ -0,0 +1,576 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "total_parameters": 49867145216,
4
+ "total_size": 199468580864
5
+ },
6
+ "weight_map": {
7
+ "lm_head.weight": "model-00044-of-00044.safetensors",
8
+ "model.embed_tokens.weight": "model-00001-of-00044.safetensors",
9
+ "model.layers.0.input_layernorm.weight": "model-00001-of-00044.safetensors",
10
+ "model.layers.0.mlp.down_proj.weight": "model-00002-of-00044.safetensors",
11
+ "model.layers.0.mlp.gate_proj.weight": "model-00002-of-00044.safetensors",
12
+ "model.layers.0.mlp.up_proj.weight": "model-00002-of-00044.safetensors",
13
+ "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00044.safetensors",
14
+ "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00044.safetensors",
15
+ "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00044.safetensors",
16
+ "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00044.safetensors",
17
+ "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00044.safetensors",
18
+ "model.layers.1.input_layernorm.weight": "model-00002-of-00044.safetensors",
19
+ "model.layers.1.mlp.down_proj.weight": "model-00002-of-00044.safetensors",
20
+ "model.layers.1.mlp.gate_proj.weight": "model-00002-of-00044.safetensors",
21
+ "model.layers.1.mlp.up_proj.weight": "model-00002-of-00044.safetensors",
22
+ "model.layers.1.post_attention_layernorm.weight": "model-00002-of-00044.safetensors",
23
+ "model.layers.1.self_attn.k_proj.weight": "model-00002-of-00044.safetensors",
24
+ "model.layers.1.self_attn.o_proj.weight": "model-00002-of-00044.safetensors",
25
+ "model.layers.1.self_attn.q_proj.weight": "model-00002-of-00044.safetensors",
26
+ "model.layers.1.self_attn.v_proj.weight": "model-00002-of-00044.safetensors",
27
+ "model.layers.10.input_layernorm.weight": "model-00007-of-00044.safetensors",
28
+ "model.layers.10.mlp.down_proj.weight": "model-00008-of-00044.safetensors",
29
+ "model.layers.10.mlp.gate_proj.weight": "model-00008-of-00044.safetensors",
30
+ "model.layers.10.mlp.up_proj.weight": "model-00008-of-00044.safetensors",
31
+ "model.layers.10.post_attention_layernorm.weight": "model-00007-of-00044.safetensors",
32
+ "model.layers.10.self_attn.k_proj.weight": "model-00007-of-00044.safetensors",
33
+ "model.layers.10.self_attn.o_proj.weight": "model-00007-of-00044.safetensors",
34
+ "model.layers.10.self_attn.q_proj.weight": "model-00007-of-00044.safetensors",
35
+ "model.layers.10.self_attn.v_proj.weight": "model-00007-of-00044.safetensors",
36
+ "model.layers.11.mlp.down_proj.weight": "model-00008-of-00044.safetensors",
37
+ "model.layers.11.mlp.gate_proj.weight": "model-00008-of-00044.safetensors",
38
+ "model.layers.11.mlp.up_proj.weight": "model-00008-of-00044.safetensors",
39
+ "model.layers.11.post_attention_layernorm.weight": "model-00008-of-00044.safetensors",
40
+ "model.layers.12.input_layernorm.weight": "model-00008-of-00044.safetensors",
41
+ "model.layers.12.mlp.down_proj.weight": "model-00009-of-00044.safetensors",
42
+ "model.layers.12.mlp.gate_proj.weight": "model-00009-of-00044.safetensors",
43
+ "model.layers.12.mlp.up_proj.weight": "model-00009-of-00044.safetensors",
44
+ "model.layers.12.post_attention_layernorm.weight": "model-00009-of-00044.safetensors",
45
+ "model.layers.12.self_attn.k_proj.weight": "model-00008-of-00044.safetensors",
46
+ "model.layers.12.self_attn.o_proj.weight": "model-00009-of-00044.safetensors",
47
+ "model.layers.12.self_attn.q_proj.weight": "model-00008-of-00044.safetensors",
48
+ "model.layers.12.self_attn.v_proj.weight": "model-00008-of-00044.safetensors",
49
+ "model.layers.13.input_layernorm.weight": "model-00009-of-00044.safetensors",
50
+ "model.layers.13.mlp.down_proj.weight": "model-00010-of-00044.safetensors",
51
+ "model.layers.13.mlp.gate_proj.weight": "model-00009-of-00044.safetensors",
52
+ "model.layers.13.mlp.up_proj.weight": "model-00010-of-00044.safetensors",
53
+ "model.layers.13.post_attention_layernorm.weight": "model-00009-of-00044.safetensors",
54
+ "model.layers.13.self_attn.k_proj.weight": "model-00009-of-00044.safetensors",
55
+ "model.layers.13.self_attn.o_proj.weight": "model-00009-of-00044.safetensors",
56
+ "model.layers.13.self_attn.q_proj.weight": "model-00009-of-00044.safetensors",
57
+ "model.layers.13.self_attn.v_proj.weight": "model-00009-of-00044.safetensors",
58
+ "model.layers.14.input_layernorm.weight": "model-00010-of-00044.safetensors",
59
+ "model.layers.14.mlp.down_proj.weight": "model-00011-of-00044.safetensors",
60
+ "model.layers.14.mlp.gate_proj.weight": "model-00010-of-00044.safetensors",
61
+ "model.layers.14.mlp.up_proj.weight": "model-00010-of-00044.safetensors",
62
+ "model.layers.14.post_attention_layernorm.weight": "model-00010-of-00044.safetensors",
63
+ "model.layers.14.self_attn.k_proj.weight": "model-00010-of-00044.safetensors",
64
+ "model.layers.14.self_attn.o_proj.weight": "model-00010-of-00044.safetensors",
65
+ "model.layers.14.self_attn.q_proj.weight": "model-00010-of-00044.safetensors",
66
+ "model.layers.14.self_attn.v_proj.weight": "model-00010-of-00044.safetensors",
67
+ "model.layers.15.input_layernorm.weight": "model-00011-of-00044.safetensors",
68
+ "model.layers.15.mlp.down_proj.weight": "model-00011-of-00044.safetensors",
69
+ "model.layers.15.mlp.gate_proj.weight": "model-00011-of-00044.safetensors",
70
+ "model.layers.15.mlp.up_proj.weight": "model-00011-of-00044.safetensors",
71
+ "model.layers.15.post_attention_layernorm.weight": "model-00011-of-00044.safetensors",
72
+ "model.layers.15.self_attn.k_proj.weight": "model-00011-of-00044.safetensors",
73
+ "model.layers.15.self_attn.o_proj.weight": "model-00011-of-00044.safetensors",
74
+ "model.layers.15.self_attn.q_proj.weight": "model-00011-of-00044.safetensors",
75
+ "model.layers.15.self_attn.v_proj.weight": "model-00011-of-00044.safetensors",
76
+ "model.layers.16.input_layernorm.weight": "model-00011-of-00044.safetensors",
77
+ "model.layers.16.mlp.down_proj.weight": "model-00012-of-00044.safetensors",
78
+ "model.layers.16.mlp.gate_proj.weight": "model-00012-of-00044.safetensors",
79
+ "model.layers.16.mlp.up_proj.weight": "model-00012-of-00044.safetensors",
80
+ "model.layers.16.post_attention_layernorm.weight": "model-00011-of-00044.safetensors",
81
+ "model.layers.16.self_attn.k_proj.weight": "model-00011-of-00044.safetensors",
82
+ "model.layers.16.self_attn.o_proj.weight": "model-00011-of-00044.safetensors",
83
+ "model.layers.16.self_attn.q_proj.weight": "model-00011-of-00044.safetensors",
84
+ "model.layers.16.self_attn.v_proj.weight": "model-00011-of-00044.safetensors",
85
+ "model.layers.17.input_layernorm.weight": "model-00012-of-00044.safetensors",
86
+ "model.layers.17.mlp.down_proj.weight": "model-00013-of-00044.safetensors",
87
+ "model.layers.17.mlp.gate_proj.weight": "model-00012-of-00044.safetensors",
88
+ "model.layers.17.mlp.up_proj.weight": "model-00013-of-00044.safetensors",
89
+ "model.layers.17.post_attention_layernorm.weight": "model-00012-of-00044.safetensors",
90
+ "model.layers.17.self_attn.k_proj.weight": "model-00012-of-00044.safetensors",
91
+ "model.layers.17.self_attn.o_proj.weight": "model-00012-of-00044.safetensors",
92
+ "model.layers.17.self_attn.q_proj.weight": "model-00012-of-00044.safetensors",
93
+ "model.layers.17.self_attn.v_proj.weight": "model-00012-of-00044.safetensors",
94
+ "model.layers.18.input_layernorm.weight": "model-00013-of-00044.safetensors",
95
+ "model.layers.18.mlp.down_proj.weight": "model-00014-of-00044.safetensors",
96
+ "model.layers.18.mlp.gate_proj.weight": "model-00013-of-00044.safetensors",
97
+ "model.layers.18.mlp.up_proj.weight": "model-00013-of-00044.safetensors",
98
+ "model.layers.18.post_attention_layernorm.weight": "model-00013-of-00044.safetensors",
99
+ "model.layers.18.self_attn.k_proj.weight": "model-00013-of-00044.safetensors",
100
+ "model.layers.18.self_attn.o_proj.weight": "model-00013-of-00044.safetensors",
101
+ "model.layers.18.self_attn.q_proj.weight": "model-00013-of-00044.safetensors",
102
+ "model.layers.18.self_attn.v_proj.weight": "model-00013-of-00044.safetensors",
103
+ "model.layers.19.input_layernorm.weight": "model-00014-of-00044.safetensors",
104
+ "model.layers.19.mlp.down_proj.weight": "model-00014-of-00044.safetensors",
105
+ "model.layers.19.mlp.gate_proj.weight": "model-00014-of-00044.safetensors",
106
+ "model.layers.19.mlp.up_proj.weight": "model-00014-of-00044.safetensors",
107
+ "model.layers.19.post_attention_layernorm.weight": "model-00014-of-00044.safetensors",
108
+ "model.layers.19.self_attn.k_proj.weight": "model-00014-of-00044.safetensors",
109
+ "model.layers.19.self_attn.o_proj.weight": "model-00014-of-00044.safetensors",
110
+ "model.layers.19.self_attn.q_proj.weight": "model-00014-of-00044.safetensors",
111
+ "model.layers.19.self_attn.v_proj.weight": "model-00014-of-00044.safetensors",
112
+ "model.layers.2.input_layernorm.weight": "model-00002-of-00044.safetensors",
113
+ "model.layers.2.mlp.down_proj.weight": "model-00003-of-00044.safetensors",
114
+ "model.layers.2.mlp.gate_proj.weight": "model-00003-of-00044.safetensors",
115
+ "model.layers.2.mlp.up_proj.weight": "model-00003-of-00044.safetensors",
116
+ "model.layers.2.post_attention_layernorm.weight": "model-00003-of-00044.safetensors",
117
+ "model.layers.2.self_attn.k_proj.weight": "model-00003-of-00044.safetensors",
118
+ "model.layers.2.self_attn.o_proj.weight": "model-00003-of-00044.safetensors",
119
+ "model.layers.2.self_attn.q_proj.weight": "model-00003-of-00044.safetensors",
120
+ "model.layers.2.self_attn.v_proj.weight": "model-00003-of-00044.safetensors",
121
+ "model.layers.20.input_layernorm.weight": "model-00014-of-00044.safetensors",
122
+ "model.layers.20.mlp.down_proj.weight": "model-00015-of-00044.safetensors",
123
+ "model.layers.20.mlp.gate_proj.weight": "model-00015-of-00044.safetensors",
124
+ "model.layers.20.mlp.up_proj.weight": "model-00015-of-00044.safetensors",
125
+ "model.layers.20.post_attention_layernorm.weight": "model-00014-of-00044.safetensors",
126
+ "model.layers.20.self_attn.k_proj.weight": "model-00014-of-00044.safetensors",
127
+ "model.layers.20.self_attn.o_proj.weight": "model-00014-of-00044.safetensors",
128
+ "model.layers.20.self_attn.q_proj.weight": "model-00014-of-00044.safetensors",
129
+ "model.layers.20.self_attn.v_proj.weight": "model-00014-of-00044.safetensors",
130
+ "model.layers.21.input_layernorm.weight": "model-00015-of-00044.safetensors",
131
+ "model.layers.21.mlp.down_proj.weight": "model-00016-of-00044.safetensors",
132
+ "model.layers.21.mlp.gate_proj.weight": "model-00015-of-00044.safetensors",
133
+ "model.layers.21.mlp.up_proj.weight": "model-00016-of-00044.safetensors",
134
+ "model.layers.21.post_attention_layernorm.weight": "model-00015-of-00044.safetensors",
135
+ "model.layers.21.self_attn.k_proj.weight": "model-00015-of-00044.safetensors",
136
+ "model.layers.21.self_attn.o_proj.weight": "model-00015-of-00044.safetensors",
137
+ "model.layers.21.self_attn.q_proj.weight": "model-00015-of-00044.safetensors",
138
+ "model.layers.21.self_attn.v_proj.weight": "model-00015-of-00044.safetensors",
139
+ "model.layers.22.input_layernorm.weight": "model-00016-of-00044.safetensors",
140
+ "model.layers.22.mlp.down_proj.weight": "model-00017-of-00044.safetensors",
141
+ "model.layers.22.mlp.gate_proj.weight": "model-00016-of-00044.safetensors",
142
+ "model.layers.22.mlp.up_proj.weight": "model-00016-of-00044.safetensors",
143
+ "model.layers.22.post_attention_layernorm.weight": "model-00016-of-00044.safetensors",
144
+ "model.layers.22.self_attn.k_proj.weight": "model-00016-of-00044.safetensors",
145
+ "model.layers.22.self_attn.o_proj.weight": "model-00016-of-00044.safetensors",
146
+ "model.layers.22.self_attn.q_proj.weight": "model-00016-of-00044.safetensors",
147
+ "model.layers.22.self_attn.v_proj.weight": "model-00016-of-00044.safetensors",
148
+ "model.layers.23.input_layernorm.weight": "model-00017-of-00044.safetensors",
149
+ "model.layers.23.mlp.down_proj.weight": "model-00017-of-00044.safetensors",
150
+ "model.layers.23.mlp.gate_proj.weight": "model-00017-of-00044.safetensors",
151
+ "model.layers.23.mlp.up_proj.weight": "model-00017-of-00044.safetensors",
152
+ "model.layers.23.post_attention_layernorm.weight": "model-00017-of-00044.safetensors",
153
+ "model.layers.23.self_attn.k_proj.weight": "model-00017-of-00044.safetensors",
154
+ "model.layers.23.self_attn.o_proj.weight": "model-00017-of-00044.safetensors",
155
+ "model.layers.23.self_attn.q_proj.weight": "model-00017-of-00044.safetensors",
156
+ "model.layers.23.self_attn.v_proj.weight": "model-00017-of-00044.safetensors",
157
+ "model.layers.24.input_layernorm.weight": "model-00017-of-00044.safetensors",
158
+ "model.layers.24.mlp.down_proj.weight": "model-00018-of-00044.safetensors",
159
+ "model.layers.24.mlp.gate_proj.weight": "model-00018-of-00044.safetensors",
160
+ "model.layers.24.mlp.up_proj.weight": "model-00018-of-00044.safetensors",
161
+ "model.layers.24.post_attention_layernorm.weight": "model-00017-of-00044.safetensors",
162
+ "model.layers.24.self_attn.k_proj.weight": "model-00017-of-00044.safetensors",
163
+ "model.layers.24.self_attn.o_proj.weight": "model-00017-of-00044.safetensors",
164
+ "model.layers.24.self_attn.q_proj.weight": "model-00017-of-00044.safetensors",
165
+ "model.layers.24.self_attn.v_proj.weight": "model-00017-of-00044.safetensors",
166
+ "model.layers.25.input_layernorm.weight": "model-00018-of-00044.safetensors",
167
+ "model.layers.25.mlp.down_proj.weight": "model-00019-of-00044.safetensors",
168
+ "model.layers.25.mlp.gate_proj.weight": "model-00018-of-00044.safetensors",
169
+ "model.layers.25.mlp.up_proj.weight": "model-00019-of-00044.safetensors",
170
+ "model.layers.25.post_attention_layernorm.weight": "model-00018-of-00044.safetensors",
171
+ "model.layers.25.self_attn.k_proj.weight": "model-00018-of-00044.safetensors",
172
+ "model.layers.25.self_attn.o_proj.weight": "model-00018-of-00044.safetensors",
173
+ "model.layers.25.self_attn.q_proj.weight": "model-00018-of-00044.safetensors",
174
+ "model.layers.25.self_attn.v_proj.weight": "model-00018-of-00044.safetensors",
175
+ "model.layers.26.input_layernorm.weight": "model-00019-of-00044.safetensors",
176
+ "model.layers.26.mlp.down_proj.weight": "model-00020-of-00044.safetensors",
177
+ "model.layers.26.mlp.gate_proj.weight": "model-00019-of-00044.safetensors",
178
+ "model.layers.26.mlp.up_proj.weight": "model-00019-of-00044.safetensors",
179
+ "model.layers.26.post_attention_layernorm.weight": "model-00019-of-00044.safetensors",
180
+ "model.layers.26.self_attn.k_proj.weight": "model-00019-of-00044.safetensors",
181
+ "model.layers.26.self_attn.o_proj.weight": "model-00019-of-00044.safetensors",
182
+ "model.layers.26.self_attn.q_proj.weight": "model-00019-of-00044.safetensors",
183
+ "model.layers.26.self_attn.v_proj.weight": "model-00019-of-00044.safetensors",
184
+ "model.layers.27.input_layernorm.weight": "model-00020-of-00044.safetensors",
185
+ "model.layers.27.mlp.down_proj.weight": "model-00020-of-00044.safetensors",
186
+ "model.layers.27.mlp.gate_proj.weight": "model-00020-of-00044.safetensors",
187
+ "model.layers.27.mlp.up_proj.weight": "model-00020-of-00044.safetensors",
188
+ "model.layers.27.post_attention_layernorm.weight": "model-00020-of-00044.safetensors",
189
+ "model.layers.27.self_attn.k_proj.weight": "model-00020-of-00044.safetensors",
190
+ "model.layers.27.self_attn.o_proj.weight": "model-00020-of-00044.safetensors",
191
+ "model.layers.27.self_attn.q_proj.weight": "model-00020-of-00044.safetensors",
192
+ "model.layers.27.self_attn.v_proj.weight": "model-00020-of-00044.safetensors",
193
+ "model.layers.28.input_layernorm.weight": "model-00020-of-00044.safetensors",
194
+ "model.layers.28.mlp.down_proj.weight": "model-00021-of-00044.safetensors",
195
+ "model.layers.28.mlp.gate_proj.weight": "model-00021-of-00044.safetensors",
196
+ "model.layers.28.mlp.up_proj.weight": "model-00021-of-00044.safetensors",
197
+ "model.layers.28.post_attention_layernorm.weight": "model-00020-of-00044.safetensors",
198
+ "model.layers.28.self_attn.k_proj.weight": "model-00020-of-00044.safetensors",
199
+ "model.layers.28.self_attn.o_proj.weight": "model-00020-of-00044.safetensors",
200
+ "model.layers.28.self_attn.q_proj.weight": "model-00020-of-00044.safetensors",
201
+ "model.layers.28.self_attn.v_proj.weight": "model-00020-of-00044.safetensors",
202
+ "model.layers.29.input_layernorm.weight": "model-00021-of-00044.safetensors",
203
+ "model.layers.29.mlp.down_proj.weight": "model-00022-of-00044.safetensors",
204
+ "model.layers.29.mlp.gate_proj.weight": "model-00021-of-00044.safetensors",
205
+ "model.layers.29.mlp.up_proj.weight": "model-00022-of-00044.safetensors",
206
+ "model.layers.29.post_attention_layernorm.weight": "model-00021-of-00044.safetensors",
207
+ "model.layers.29.self_attn.k_proj.weight": "model-00021-of-00044.safetensors",
208
+ "model.layers.29.self_attn.o_proj.weight": "model-00021-of-00044.safetensors",
209
+ "model.layers.29.self_attn.q_proj.weight": "model-00021-of-00044.safetensors",
210
+ "model.layers.29.self_attn.v_proj.weight": "model-00021-of-00044.safetensors",
211
+ "model.layers.3.input_layernorm.weight": "model-00003-of-00044.safetensors",
212
+ "model.layers.3.mlp.down_proj.weight": "model-00004-of-00044.safetensors",
213
+ "model.layers.3.mlp.gate_proj.weight": "model-00003-of-00044.safetensors",
214
+ "model.layers.3.mlp.up_proj.weight": "model-00004-of-00044.safetensors",
215
+ "model.layers.3.post_attention_layernorm.weight": "model-00003-of-00044.safetensors",
216
+ "model.layers.3.self_attn.k_proj.weight": "model-00003-of-00044.safetensors",
217
+ "model.layers.3.self_attn.o_proj.weight": "model-00003-of-00044.safetensors",
218
+ "model.layers.3.self_attn.q_proj.weight": "model-00003-of-00044.safetensors",
219
+ "model.layers.3.self_attn.v_proj.weight": "model-00003-of-00044.safetensors",
220
+ "model.layers.30.input_layernorm.weight": "model-00022-of-00044.safetensors",
221
+ "model.layers.30.mlp.down_proj.weight": "model-00023-of-00044.safetensors",
222
+ "model.layers.30.mlp.gate_proj.weight": "model-00022-of-00044.safetensors",
223
+ "model.layers.30.mlp.up_proj.weight": "model-00022-of-00044.safetensors",
224
+ "model.layers.30.post_attention_layernorm.weight": "model-00022-of-00044.safetensors",
225
+ "model.layers.30.self_attn.k_proj.weight": "model-00022-of-00044.safetensors",
226
+ "model.layers.30.self_attn.o_proj.weight": "model-00022-of-00044.safetensors",
227
+ "model.layers.30.self_attn.q_proj.weight": "model-00022-of-00044.safetensors",
228
+ "model.layers.30.self_attn.v_proj.weight": "model-00022-of-00044.safetensors",
229
+ "model.layers.31.input_layernorm.weight": "model-00023-of-00044.safetensors",
230
+ "model.layers.31.mlp.down_proj.weight": "model-00023-of-00044.safetensors",
231
+ "model.layers.31.mlp.gate_proj.weight": "model-00023-of-00044.safetensors",
232
+ "model.layers.31.mlp.up_proj.weight": "model-00023-of-00044.safetensors",
233
+ "model.layers.31.post_attention_layernorm.weight": "model-00023-of-00044.safetensors",
234
+ "model.layers.31.self_attn.k_proj.weight": "model-00023-of-00044.safetensors",
235
+ "model.layers.31.self_attn.o_proj.weight": "model-00023-of-00044.safetensors",
236
+ "model.layers.31.self_attn.q_proj.weight": "model-00023-of-00044.safetensors",
237
+ "model.layers.31.self_attn.v_proj.weight": "model-00023-of-00044.safetensors",
238
+ "model.layers.32.input_layernorm.weight": "model-00023-of-00044.safetensors",
239
+ "model.layers.32.mlp.down_proj.weight": "model-00024-of-00044.safetensors",
240
+ "model.layers.32.mlp.gate_proj.weight": "model-00024-of-00044.safetensors",
241
+ "model.layers.32.mlp.up_proj.weight": "model-00024-of-00044.safetensors",
242
+ "model.layers.32.post_attention_layernorm.weight": "model-00023-of-00044.safetensors",
243
+ "model.layers.32.self_attn.k_proj.weight": "model-00023-of-00044.safetensors",
244
+ "model.layers.32.self_attn.o_proj.weight": "model-00023-of-00044.safetensors",
245
+ "model.layers.32.self_attn.q_proj.weight": "model-00023-of-00044.safetensors",
246
+ "model.layers.32.self_attn.v_proj.weight": "model-00023-of-00044.safetensors",
247
+ "model.layers.33.input_layernorm.weight": "model-00024-of-00044.safetensors",
248
+ "model.layers.33.mlp.down_proj.weight": "model-00025-of-00044.safetensors",
249
+ "model.layers.33.mlp.gate_proj.weight": "model-00024-of-00044.safetensors",
250
+ "model.layers.33.mlp.up_proj.weight": "model-00025-of-00044.safetensors",
251
+ "model.layers.33.post_attention_layernorm.weight": "model-00024-of-00044.safetensors",
252
+ "model.layers.33.self_attn.k_proj.weight": "model-00024-of-00044.safetensors",
253
+ "model.layers.33.self_attn.o_proj.weight": "model-00024-of-00044.safetensors",
254
+ "model.layers.33.self_attn.q_proj.weight": "model-00024-of-00044.safetensors",
255
+ "model.layers.33.self_attn.v_proj.weight": "model-00024-of-00044.safetensors",
256
+ "model.layers.34.input_layernorm.weight": "model-00025-of-00044.safetensors",
257
+ "model.layers.34.mlp.down_proj.weight": "model-00026-of-00044.safetensors",
258
+ "model.layers.34.mlp.gate_proj.weight": "model-00025-of-00044.safetensors",
259
+ "model.layers.34.mlp.up_proj.weight": "model-00025-of-00044.safetensors",
260
+ "model.layers.34.post_attention_layernorm.weight": "model-00025-of-00044.safetensors",
261
+ "model.layers.34.self_attn.k_proj.weight": "model-00025-of-00044.safetensors",
262
+ "model.layers.34.self_attn.o_proj.weight": "model-00025-of-00044.safetensors",
263
+ "model.layers.34.self_attn.q_proj.weight": "model-00025-of-00044.safetensors",
264
+ "model.layers.34.self_attn.v_proj.weight": "model-00025-of-00044.safetensors",
265
+ "model.layers.35.input_layernorm.weight": "model-00026-of-00044.safetensors",
266
+ "model.layers.35.mlp.down_proj.weight": "model-00026-of-00044.safetensors",
267
+ "model.layers.35.mlp.gate_proj.weight": "model-00026-of-00044.safetensors",
268
+ "model.layers.35.mlp.up_proj.weight": "model-00026-of-00044.safetensors",
269
+ "model.layers.35.post_attention_layernorm.weight": "model-00026-of-00044.safetensors",
270
+ "model.layers.35.self_attn.k_proj.weight": "model-00026-of-00044.safetensors",
271
+ "model.layers.35.self_attn.o_proj.weight": "model-00026-of-00044.safetensors",
272
+ "model.layers.35.self_attn.q_proj.weight": "model-00026-of-00044.safetensors",
273
+ "model.layers.35.self_attn.v_proj.weight": "model-00026-of-00044.safetensors",
274
+ "model.layers.36.input_layernorm.weight": "model-00026-of-00044.safetensors",
275
+ "model.layers.36.mlp.down_proj.weight": "model-00027-of-00044.safetensors",
276
+ "model.layers.36.mlp.gate_proj.weight": "model-00027-of-00044.safetensors",
277
+ "model.layers.36.mlp.up_proj.weight": "model-00027-of-00044.safetensors",
278
+ "model.layers.36.post_attention_layernorm.weight": "model-00026-of-00044.safetensors",
279
+ "model.layers.36.self_attn.k_proj.weight": "model-00026-of-00044.safetensors",
280
+ "model.layers.36.self_attn.o_proj.weight": "model-00026-of-00044.safetensors",
281
+ "model.layers.36.self_attn.q_proj.weight": "model-00026-of-00044.safetensors",
282
+ "model.layers.36.self_attn.v_proj.weight": "model-00026-of-00044.safetensors",
283
+ "model.layers.37.input_layernorm.weight": "model-00027-of-00044.safetensors",
284
+ "model.layers.37.mlp.down_proj.weight": "model-00028-of-00044.safetensors",
285
+ "model.layers.37.mlp.gate_proj.weight": "model-00027-of-00044.safetensors",
286
+ "model.layers.37.mlp.up_proj.weight": "model-00028-of-00044.safetensors",
287
+ "model.layers.37.post_attention_layernorm.weight": "model-00027-of-00044.safetensors",
288
+ "model.layers.37.self_attn.k_proj.weight": "model-00027-of-00044.safetensors",
289
+ "model.layers.37.self_attn.o_proj.weight": "model-00027-of-00044.safetensors",
290
+ "model.layers.37.self_attn.q_proj.weight": "model-00027-of-00044.safetensors",
291
+ "model.layers.37.self_attn.v_proj.weight": "model-00027-of-00044.safetensors",
292
+ "model.layers.38.input_layernorm.weight": "model-00028-of-00044.safetensors",
293
+ "model.layers.38.mlp.down_proj.weight": "model-00029-of-00044.safetensors",
294
+ "model.layers.38.mlp.gate_proj.weight": "model-00028-of-00044.safetensors",
295
+ "model.layers.38.mlp.up_proj.weight": "model-00028-of-00044.safetensors",
296
+ "model.layers.38.post_attention_layernorm.weight": "model-00028-of-00044.safetensors",
297
+ "model.layers.38.self_attn.k_proj.weight": "model-00028-of-00044.safetensors",
298
+ "model.layers.38.self_attn.o_proj.weight": "model-00028-of-00044.safetensors",
299
+ "model.layers.38.self_attn.q_proj.weight": "model-00028-of-00044.safetensors",
300
+ "model.layers.38.self_attn.v_proj.weight": "model-00028-of-00044.safetensors",
301
+ "model.layers.39.input_layernorm.weight": "model-00029-of-00044.safetensors",
302
+ "model.layers.39.mlp.down_proj.weight": "model-00029-of-00044.safetensors",
303
+ "model.layers.39.mlp.gate_proj.weight": "model-00029-of-00044.safetensors",
304
+ "model.layers.39.mlp.up_proj.weight": "model-00029-of-00044.safetensors",
305
+ "model.layers.39.post_attention_layernorm.weight": "model-00029-of-00044.safetensors",
306
+ "model.layers.39.self_attn.k_proj.weight": "model-00029-of-00044.safetensors",
307
+ "model.layers.39.self_attn.o_proj.weight": "model-00029-of-00044.safetensors",
308
+ "model.layers.39.self_attn.q_proj.weight": "model-00029-of-00044.safetensors",
309
+ "model.layers.39.self_attn.v_proj.weight": "model-00029-of-00044.safetensors",
310
+ "model.layers.4.input_layernorm.weight": "model-00004-of-00044.safetensors",
311
+ "model.layers.4.mlp.down_proj.weight": "model-00005-of-00044.safetensors",
312
+ "model.layers.4.mlp.gate_proj.weight": "model-00004-of-00044.safetensors",
313
+ "model.layers.4.mlp.up_proj.weight": "model-00004-of-00044.safetensors",
314
+ "model.layers.4.post_attention_layernorm.weight": "model-00004-of-00044.safetensors",
315
+ "model.layers.4.self_attn.k_proj.weight": "model-00004-of-00044.safetensors",
316
+ "model.layers.4.self_attn.o_proj.weight": "model-00004-of-00044.safetensors",
317
+ "model.layers.4.self_attn.q_proj.weight": "model-00004-of-00044.safetensors",
318
+ "model.layers.4.self_attn.v_proj.weight": "model-00004-of-00044.safetensors",
319
+ "model.layers.40.input_layernorm.weight": "model-00029-of-00044.safetensors",
320
+ "model.layers.40.mlp.down_proj.weight": "model-00030-of-00044.safetensors",
321
+ "model.layers.40.mlp.gate_proj.weight": "model-00030-of-00044.safetensors",
322
+ "model.layers.40.mlp.up_proj.weight": "model-00030-of-00044.safetensors",
323
+ "model.layers.40.post_attention_layernorm.weight": "model-00029-of-00044.safetensors",
324
+ "model.layers.40.self_attn.k_proj.weight": "model-00029-of-00044.safetensors",
325
+ "model.layers.40.self_attn.o_proj.weight": "model-00029-of-00044.safetensors",
326
+ "model.layers.40.self_attn.q_proj.weight": "model-00029-of-00044.safetensors",
327
+ "model.layers.40.self_attn.v_proj.weight": "model-00029-of-00044.safetensors",
328
+ "model.layers.41.input_layernorm.weight": "model-00030-of-00044.safetensors",
329
+ "model.layers.41.mlp.down_proj.weight": "model-00031-of-00044.safetensors",
330
+ "model.layers.41.mlp.gate_proj.weight": "model-00030-of-00044.safetensors",
331
+ "model.layers.41.mlp.up_proj.weight": "model-00031-of-00044.safetensors",
332
+ "model.layers.41.post_attention_layernorm.weight": "model-00030-of-00044.safetensors",
333
+ "model.layers.41.self_attn.k_proj.weight": "model-00030-of-00044.safetensors",
334
+ "model.layers.41.self_attn.o_proj.weight": "model-00030-of-00044.safetensors",
335
+ "model.layers.41.self_attn.q_proj.weight": "model-00030-of-00044.safetensors",
336
+ "model.layers.41.self_attn.v_proj.weight": "model-00030-of-00044.safetensors",
337
+ "model.layers.42.mlp.down_proj.weight": "model-00031-of-00044.safetensors",
338
+ "model.layers.42.mlp.gate_proj.weight": "model-00031-of-00044.safetensors",
339
+ "model.layers.42.mlp.up_proj.weight": "model-00031-of-00044.safetensors",
340
+ "model.layers.42.post_attention_layernorm.weight": "model-00031-of-00044.safetensors",
341
+ "model.layers.43.mlp.down_proj.weight": "model-00031-of-00044.safetensors",
342
+ "model.layers.43.mlp.gate_proj.weight": "model-00031-of-00044.safetensors",
343
+ "model.layers.43.mlp.up_proj.weight": "model-00031-of-00044.safetensors",
344
+ "model.layers.43.post_attention_layernorm.weight": "model-00031-of-00044.safetensors",
345
+ "model.layers.44.mlp.down_proj.weight": "model-00032-of-00044.safetensors",
346
+ "model.layers.44.mlp.gate_proj.weight": "model-00031-of-00044.safetensors",
347
+ "model.layers.44.mlp.up_proj.weight": "model-00031-of-00044.safetensors",
348
+ "model.layers.44.post_attention_layernorm.weight": "model-00031-of-00044.safetensors",
349
+ "model.layers.45.mlp.down_proj.weight": "model-00032-of-00044.safetensors",
350
+ "model.layers.45.mlp.gate_proj.weight": "model-00032-of-00044.safetensors",
351
+ "model.layers.45.mlp.up_proj.weight": "model-00032-of-00044.safetensors",
352
+ "model.layers.45.post_attention_layernorm.weight": "model-00032-of-00044.safetensors",
353
+ "model.layers.46.mlp.down_proj.weight": "model-00032-of-00044.safetensors",
354
+ "model.layers.46.mlp.gate_proj.weight": "model-00032-of-00044.safetensors",
355
+ "model.layers.46.mlp.up_proj.weight": "model-00032-of-00044.safetensors",
356
+ "model.layers.46.post_attention_layernorm.weight": "model-00032-of-00044.safetensors",
357
+ "model.layers.47.mlp.down_proj.weight": "model-00032-of-00044.safetensors",
358
+ "model.layers.47.mlp.gate_proj.weight": "model-00032-of-00044.safetensors",
359
+ "model.layers.47.mlp.up_proj.weight": "model-00032-of-00044.safetensors",
360
+ "model.layers.47.post_attention_layernorm.weight": "model-00032-of-00044.safetensors",
361
+ "model.layers.48.mlp.down_proj.weight": "model-00033-of-00044.safetensors",
362
+ "model.layers.48.mlp.gate_proj.weight": "model-00033-of-00044.safetensors",
363
+ "model.layers.48.mlp.up_proj.weight": "model-00033-of-00044.safetensors",
364
+ "model.layers.48.post_attention_layernorm.weight": "model-00032-of-00044.safetensors",
365
+ "model.layers.49.mlp.down_proj.weight": "model-00033-of-00044.safetensors",
366
+ "model.layers.49.mlp.gate_proj.weight": "model-00033-of-00044.safetensors",
367
+ "model.layers.49.mlp.up_proj.weight": "model-00033-of-00044.safetensors",
368
+ "model.layers.49.post_attention_layernorm.weight": "model-00033-of-00044.safetensors",
369
+ "model.layers.5.input_layernorm.weight": "model-00005-of-00044.safetensors",
370
+ "model.layers.5.mlp.down_proj.weight": "model-00005-of-00044.safetensors",
371
+ "model.layers.5.mlp.gate_proj.weight": "model-00005-of-00044.safetensors",
372
+ "model.layers.5.mlp.up_proj.weight": "model-00005-of-00044.safetensors",
373
+ "model.layers.5.post_attention_layernorm.weight": "model-00005-of-00044.safetensors",
374
+ "model.layers.5.self_attn.k_proj.weight": "model-00005-of-00044.safetensors",
375
+ "model.layers.5.self_attn.o_proj.weight": "model-00005-of-00044.safetensors",
376
+ "model.layers.5.self_attn.q_proj.weight": "model-00005-of-00044.safetensors",
377
+ "model.layers.5.self_attn.v_proj.weight": "model-00005-of-00044.safetensors",
378
+ "model.layers.50.mlp.down_proj.weight": "model-00033-of-00044.safetensors",
379
+ "model.layers.50.mlp.gate_proj.weight": "model-00033-of-00044.safetensors",
380
+ "model.layers.50.mlp.up_proj.weight": "model-00033-of-00044.safetensors",
381
+ "model.layers.50.post_attention_layernorm.weight": "model-00033-of-00044.safetensors",
382
+ "model.layers.51.mlp.down_proj.weight": "model-00033-of-00044.safetensors",
383
+ "model.layers.51.mlp.gate_proj.weight": "model-00033-of-00044.safetensors",
384
+ "model.layers.51.mlp.up_proj.weight": "model-00033-of-00044.safetensors",
385
+ "model.layers.51.post_attention_layernorm.weight": "model-00033-of-00044.safetensors",
386
+ "model.layers.52.input_layernorm.weight": "model-00033-of-00044.safetensors",
387
+ "model.layers.52.mlp.down_proj.weight": "model-00034-of-00044.safetensors",
388
+ "model.layers.52.mlp.gate_proj.weight": "model-00034-of-00044.safetensors",
389
+ "model.layers.52.mlp.up_proj.weight": "model-00034-of-00044.safetensors",
390
+ "model.layers.52.post_attention_layernorm.weight": "model-00033-of-00044.safetensors",
391
+ "model.layers.52.self_attn.k_proj.weight": "model-00033-of-00044.safetensors",
392
+ "model.layers.52.self_attn.o_proj.weight": "model-00033-of-00044.safetensors",
393
+ "model.layers.52.self_attn.q_proj.weight": "model-00033-of-00044.safetensors",
394
+ "model.layers.52.self_attn.v_proj.weight": "model-00033-of-00044.safetensors",
395
+ "model.layers.53.mlp.down_proj.weight": "model-00034-of-00044.safetensors",
396
+ "model.layers.53.mlp.gate_proj.weight": "model-00034-of-00044.safetensors",
397
+ "model.layers.53.mlp.up_proj.weight": "model-00034-of-00044.safetensors",
398
+ "model.layers.53.post_attention_layernorm.weight": "model-00034-of-00044.safetensors",
399
+ "model.layers.54.mlp.down_proj.weight": "model-00034-of-00044.safetensors",
400
+ "model.layers.54.mlp.gate_proj.weight": "model-00034-of-00044.safetensors",
401
+ "model.layers.54.mlp.up_proj.weight": "model-00034-of-00044.safetensors",
402
+ "model.layers.54.post_attention_layernorm.weight": "model-00034-of-00044.safetensors",
403
+ "model.layers.55.mlp.down_proj.weight": "model-00034-of-00044.safetensors",
404
+ "model.layers.55.mlp.gate_proj.weight": "model-00034-of-00044.safetensors",
405
+ "model.layers.55.mlp.up_proj.weight": "model-00034-of-00044.safetensors",
406
+ "model.layers.55.post_attention_layernorm.weight": "model-00034-of-00044.safetensors",
407
+ "model.layers.56.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
408
+ "model.layers.56.mlp.gate_proj.weight": "model-00034-of-00044.safetensors",
409
+ "model.layers.56.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
410
+ "model.layers.56.post_attention_layernorm.weight": "model-00034-of-00044.safetensors",
411
+ "model.layers.57.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
412
+ "model.layers.57.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
413
+ "model.layers.57.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
414
+ "model.layers.57.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
415
+ "model.layers.58.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
416
+ "model.layers.58.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
417
+ "model.layers.58.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
418
+ "model.layers.58.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
419
+ "model.layers.59.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
420
+ "model.layers.59.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
421
+ "model.layers.59.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
422
+ "model.layers.59.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
423
+ "model.layers.6.mlp.down_proj.weight": "model-00006-of-00044.safetensors",
424
+ "model.layers.6.mlp.gate_proj.weight": "model-00005-of-00044.safetensors",
425
+ "model.layers.6.mlp.up_proj.weight": "model-00006-of-00044.safetensors",
426
+ "model.layers.6.post_attention_layernorm.weight": "model-00005-of-00044.safetensors",
427
+ "model.layers.60.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
428
+ "model.layers.60.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
429
+ "model.layers.60.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
430
+ "model.layers.60.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
431
+ "model.layers.61.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
432
+ "model.layers.61.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
433
+ "model.layers.61.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
434
+ "model.layers.61.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
435
+ "model.layers.62.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
436
+ "model.layers.62.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
437
+ "model.layers.62.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
438
+ "model.layers.62.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
439
+ "model.layers.63.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
440
+ "model.layers.63.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
441
+ "model.layers.63.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
442
+ "model.layers.63.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
443
+ "model.layers.64.mlp.down_proj.weight": "model-00035-of-00044.safetensors",
444
+ "model.layers.64.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
445
+ "model.layers.64.mlp.up_proj.weight": "model-00035-of-00044.safetensors",
446
+ "model.layers.64.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
447
+ "model.layers.65.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
448
+ "model.layers.65.mlp.gate_proj.weight": "model-00035-of-00044.safetensors",
449
+ "model.layers.65.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
450
+ "model.layers.65.post_attention_layernorm.weight": "model-00035-of-00044.safetensors",
451
+ "model.layers.66.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
452
+ "model.layers.66.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
453
+ "model.layers.66.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
454
+ "model.layers.66.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
455
+ "model.layers.67.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
456
+ "model.layers.67.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
457
+ "model.layers.67.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
458
+ "model.layers.67.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
459
+ "model.layers.68.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
460
+ "model.layers.68.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
461
+ "model.layers.68.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
462
+ "model.layers.68.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
463
+ "model.layers.69.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
464
+ "model.layers.69.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
465
+ "model.layers.69.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
466
+ "model.layers.69.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
467
+ "model.layers.7.mlp.down_proj.weight": "model-00006-of-00044.safetensors",
468
+ "model.layers.7.mlp.gate_proj.weight": "model-00006-of-00044.safetensors",
469
+ "model.layers.7.mlp.up_proj.weight": "model-00006-of-00044.safetensors",
470
+ "model.layers.7.post_attention_layernorm.weight": "model-00006-of-00044.safetensors",
471
+ "model.layers.70.mlp.down_proj.weight": "model-00036-of-00044.safetensors",
472
+ "model.layers.70.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
473
+ "model.layers.70.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
474
+ "model.layers.70.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
475
+ "model.layers.71.input_layernorm.weight": "model-00036-of-00044.safetensors",
476
+ "model.layers.71.mlp.down_proj.weight": "model-00037-of-00044.safetensors",
477
+ "model.layers.71.mlp.gate_proj.weight": "model-00036-of-00044.safetensors",
478
+ "model.layers.71.mlp.up_proj.weight": "model-00036-of-00044.safetensors",
479
+ "model.layers.71.post_attention_layernorm.weight": "model-00036-of-00044.safetensors",
480
+ "model.layers.71.self_attn.k_proj.weight": "model-00036-of-00044.safetensors",
481
+ "model.layers.71.self_attn.o_proj.weight": "model-00036-of-00044.safetensors",
482
+ "model.layers.71.self_attn.q_proj.weight": "model-00036-of-00044.safetensors",
483
+ "model.layers.71.self_attn.v_proj.weight": "model-00036-of-00044.safetensors",
484
+ "model.layers.72.input_layernorm.weight": "model-00037-of-00044.safetensors",
485
+ "model.layers.72.mlp.down_proj.weight": "model-00037-of-00044.safetensors",
486
+ "model.layers.72.mlp.gate_proj.weight": "model-00037-of-00044.safetensors",
487
+ "model.layers.72.mlp.up_proj.weight": "model-00037-of-00044.safetensors",
488
+ "model.layers.72.post_attention_layernorm.weight": "model-00037-of-00044.safetensors",
489
+ "model.layers.72.self_attn.k_proj.weight": "model-00037-of-00044.safetensors",
490
+ "model.layers.72.self_attn.o_proj.weight": "model-00037-of-00044.safetensors",
491
+ "model.layers.72.self_attn.q_proj.weight": "model-00037-of-00044.safetensors",
492
+ "model.layers.72.self_attn.v_proj.weight": "model-00037-of-00044.safetensors",
493
+ "model.layers.73.input_layernorm.weight": "model-00037-of-00044.safetensors",
494
+ "model.layers.73.mlp.down_proj.weight": "model-00038-of-00044.safetensors",
495
+ "model.layers.73.mlp.gate_proj.weight": "model-00038-of-00044.safetensors",
496
+ "model.layers.73.mlp.up_proj.weight": "model-00038-of-00044.safetensors",
497
+ "model.layers.73.post_attention_layernorm.weight": "model-00037-of-00044.safetensors",
498
+ "model.layers.73.self_attn.k_proj.weight": "model-00037-of-00044.safetensors",
499
+ "model.layers.73.self_attn.o_proj.weight": "model-00037-of-00044.safetensors",
500
+ "model.layers.73.self_attn.q_proj.weight": "model-00037-of-00044.safetensors",
501
+ "model.layers.73.self_attn.v_proj.weight": "model-00037-of-00044.safetensors",
502
+ "model.layers.74.input_layernorm.weight": "model-00038-of-00044.safetensors",
503
+ "model.layers.74.mlp.down_proj.weight": "model-00039-of-00044.safetensors",
504
+ "model.layers.74.mlp.gate_proj.weight": "model-00038-of-00044.safetensors",
505
+ "model.layers.74.mlp.up_proj.weight": "model-00039-of-00044.safetensors",
506
+ "model.layers.74.post_attention_layernorm.weight": "model-00038-of-00044.safetensors",
507
+ "model.layers.74.self_attn.k_proj.weight": "model-00038-of-00044.safetensors",
508
+ "model.layers.74.self_attn.o_proj.weight": "model-00038-of-00044.safetensors",
509
+ "model.layers.74.self_attn.q_proj.weight": "model-00038-of-00044.safetensors",
510
+ "model.layers.74.self_attn.v_proj.weight": "model-00038-of-00044.safetensors",
511
+ "model.layers.75.input_layernorm.weight": "model-00039-of-00044.safetensors",
512
+ "model.layers.75.mlp.down_proj.weight": "model-00040-of-00044.safetensors",
513
+ "model.layers.75.mlp.gate_proj.weight": "model-00039-of-00044.safetensors",
514
+ "model.layers.75.mlp.up_proj.weight": "model-00039-of-00044.safetensors",
515
+ "model.layers.75.post_attention_layernorm.weight": "model-00039-of-00044.safetensors",
516
+ "model.layers.75.self_attn.k_proj.weight": "model-00039-of-00044.safetensors",
517
+ "model.layers.75.self_attn.o_proj.weight": "model-00039-of-00044.safetensors",
518
+ "model.layers.75.self_attn.q_proj.weight": "model-00039-of-00044.safetensors",
519
+ "model.layers.75.self_attn.v_proj.weight": "model-00039-of-00044.safetensors",
520
+ "model.layers.76.input_layernorm.weight": "model-00040-of-00044.safetensors",
521
+ "model.layers.76.mlp.down_proj.weight": "model-00040-of-00044.safetensors",
522
+ "model.layers.76.mlp.gate_proj.weight": "model-00040-of-00044.safetensors",
523
+ "model.layers.76.mlp.up_proj.weight": "model-00040-of-00044.safetensors",
524
+ "model.layers.76.post_attention_layernorm.weight": "model-00040-of-00044.safetensors",
525
+ "model.layers.76.self_attn.k_proj.weight": "model-00040-of-00044.safetensors",
526
+ "model.layers.76.self_attn.o_proj.weight": "model-00040-of-00044.safetensors",
527
+ "model.layers.76.self_attn.q_proj.weight": "model-00040-of-00044.safetensors",
528
+ "model.layers.76.self_attn.v_proj.weight": "model-00040-of-00044.safetensors",
529
+ "model.layers.77.input_layernorm.weight": "model-00040-of-00044.safetensors",
530
+ "model.layers.77.mlp.down_proj.weight": "model-00041-of-00044.safetensors",
531
+ "model.layers.77.mlp.gate_proj.weight": "model-00041-of-00044.safetensors",
532
+ "model.layers.77.mlp.up_proj.weight": "model-00041-of-00044.safetensors",
533
+ "model.layers.77.post_attention_layernorm.weight": "model-00040-of-00044.safetensors",
534
+ "model.layers.77.self_attn.k_proj.weight": "model-00040-of-00044.safetensors",
535
+ "model.layers.77.self_attn.o_proj.weight": "model-00040-of-00044.safetensors",
536
+ "model.layers.77.self_attn.q_proj.weight": "model-00040-of-00044.safetensors",
537
+ "model.layers.77.self_attn.v_proj.weight": "model-00040-of-00044.safetensors",
538
+ "model.layers.78.input_layernorm.weight": "model-00041-of-00044.safetensors",
539
+ "model.layers.78.mlp.down_proj.weight": "model-00042-of-00044.safetensors",
540
+ "model.layers.78.mlp.gate_proj.weight": "model-00041-of-00044.safetensors",
541
+ "model.layers.78.mlp.up_proj.weight": "model-00042-of-00044.safetensors",
542
+ "model.layers.78.post_attention_layernorm.weight": "model-00041-of-00044.safetensors",
543
+ "model.layers.78.self_attn.k_proj.weight": "model-00041-of-00044.safetensors",
544
+ "model.layers.78.self_attn.o_proj.weight": "model-00041-of-00044.safetensors",
545
+ "model.layers.78.self_attn.q_proj.weight": "model-00041-of-00044.safetensors",
546
+ "model.layers.78.self_attn.v_proj.weight": "model-00041-of-00044.safetensors",
547
+ "model.layers.79.input_layernorm.weight": "model-00042-of-00044.safetensors",
548
+ "model.layers.79.mlp.down_proj.weight": "model-00043-of-00044.safetensors",
549
+ "model.layers.79.mlp.gate_proj.weight": "model-00042-of-00044.safetensors",
550
+ "model.layers.79.mlp.up_proj.weight": "model-00042-of-00044.safetensors",
551
+ "model.layers.79.post_attention_layernorm.weight": "model-00042-of-00044.safetensors",
552
+ "model.layers.79.self_attn.k_proj.weight": "model-00042-of-00044.safetensors",
553
+ "model.layers.79.self_attn.o_proj.weight": "model-00042-of-00044.safetensors",
554
+ "model.layers.79.self_attn.q_proj.weight": "model-00042-of-00044.safetensors",
555
+ "model.layers.79.self_attn.v_proj.weight": "model-00042-of-00044.safetensors",
556
+ "model.layers.8.input_layernorm.weight": "model-00006-of-00044.safetensors",
557
+ "model.layers.8.mlp.down_proj.weight": "model-00007-of-00044.safetensors",
558
+ "model.layers.8.mlp.gate_proj.weight": "model-00006-of-00044.safetensors",
559
+ "model.layers.8.mlp.up_proj.weight": "model-00006-of-00044.safetensors",
560
+ "model.layers.8.post_attention_layernorm.weight": "model-00006-of-00044.safetensors",
561
+ "model.layers.8.self_attn.k_proj.weight": "model-00006-of-00044.safetensors",
562
+ "model.layers.8.self_attn.o_proj.weight": "model-00006-of-00044.safetensors",
563
+ "model.layers.8.self_attn.q_proj.weight": "model-00006-of-00044.safetensors",
564
+ "model.layers.8.self_attn.v_proj.weight": "model-00006-of-00044.safetensors",
565
+ "model.layers.9.input_layernorm.weight": "model-00007-of-00044.safetensors",
566
+ "model.layers.9.mlp.down_proj.weight": "model-00007-of-00044.safetensors",
567
+ "model.layers.9.mlp.gate_proj.weight": "model-00007-of-00044.safetensors",
568
+ "model.layers.9.mlp.up_proj.weight": "model-00007-of-00044.safetensors",
569
+ "model.layers.9.post_attention_layernorm.weight": "model-00007-of-00044.safetensors",
570
+ "model.layers.9.self_attn.k_proj.weight": "model-00007-of-00044.safetensors",
571
+ "model.layers.9.self_attn.o_proj.weight": "model-00007-of-00044.safetensors",
572
+ "model.layers.9.self_attn.q_proj.weight": "model-00007-of-00044.safetensors",
573
+ "model.layers.9.self_attn.v_proj.weight": "model-00007-of-00044.safetensors",
574
+ "model.norm.weight": "model-00043-of-00044.safetensors"
575
+ }
576
+ }