qgallouedec HF Staff commited on
Commit
5e76121
·
verified ·
1 Parent(s): 9d893c8

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
chat_template.jinja ADDED
@@ -0,0 +1,137 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- set effort_map = {"none": 0.0, "minimal": 0.1, "low": 0.2, "medium": 0.7, "high": 0.9, "xhigh": 0.99, "max": 0.99} -%}
2
+ {%- set role_token = {"user": "<|message_user|>", "assistant": "<|message_model|>", "system": "<|message_system|>", "tool": "<|message_tool|>"} -%}
3
+
4
+ {%- macro emit_thinking_effort() -%}
5
+ {%- set eff = reasoning_effort if reasoning_effort is defined and reasoning_effort is not none else 0.9 -%}
6
+ {%- if eff is string -%}
7
+ {%- set key = eff | trim -%}
8
+ {%- if key not in effort_map -%}
9
+ {{- raise_exception("Unknown reasoning_effort: " ~ eff) -}}
10
+ {%- endif -%}
11
+ {%- set num = effort_map[key] -%}
12
+ {%- else -%}
13
+ {%- set num = eff | float -%}
14
+ {%- endif -%}
15
+ {%- if num < 0.0 or num > 0.99 -%}
16
+ {{- raise_exception("reasoning_effort must be in [0.0, 0.99]") -}}
17
+ {%- endif -%}
18
+ {{- "<|message_system|><|content_text|>Thinking effort level: " -}}
19
+ {%- if num == 0.0 -%}0{%- else -%}{{ num }}{%- endif -%}
20
+ {{- "<|end_message|>" -}}
21
+ {%- endmacro -%}
22
+
23
+ {%- if tools -%}
24
+ {%- set tool_state = namespace(specs=[]) -%}
25
+ {%- for tool in tools -%}
26
+ {%- set fn = tool.function if tool.function is defined else tool -%}
27
+ {%- set spec = {
28
+ "description": (fn.description if fn.description is defined and fn.description else ""),
29
+ "name": fn.name,
30
+ "parameters": (fn.parameters if fn.parameters is defined and fn.parameters else {}),
31
+ "type": (tool.type if tool.type is defined and tool.type else "function"),
32
+ } -%}
33
+ {%- set tool_state.specs = tool_state.specs + [spec] -%}
34
+ {%- endfor -%}
35
+ {{- "<|message_system|>tool_declare<|content_xml|>" -}}
36
+ {{- tool_state.specs | tojson(sort_keys=true, separators=(",", ":")) -}}
37
+ {{- "<|end_message|>" -}}
38
+ {%- endif -%}
39
+
40
+ {%- set state = namespace(effort_emitted=false) -%}
41
+ {%- for message in messages -%}
42
+ {%- if message.role not in role_token -%}
43
+ {{- raise_exception("Unknown message role: " ~ message.role) -}}
44
+ {%- endif -%}
45
+ {%- if not state.effort_emitted and message.role != "system" -%}
46
+ {{- emit_thinking_effort() -}}
47
+ {%- set state.effort_emitted = true -%}
48
+ {%- endif -%}
49
+
50
+ {%- set header = namespace(text=role_token[message.role]) -%}
51
+ {%- if message.role == "tool" -%}
52
+ {%- if message.name is defined and message.name -%}
53
+ {%- set header.text = header.text ~ message.name -%}
54
+ {%- elif message.tool_call_id is defined and message.tool_call_id -%}
55
+ {%- for prev in messages -%}
56
+ {%- if prev.role == "assistant" and prev.tool_calls -%}
57
+ {%- for tc in prev.tool_calls -%}
58
+ {%- if tc.id is defined and tc.id == message.tool_call_id and tc.function.name is defined -%}
59
+ {%- set header.text = header.text ~ tc.function.name -%}
60
+ {%- endif -%}
61
+ {%- endfor -%}
62
+ {%- endif -%}
63
+ {%- endfor -%}
64
+ {%- endif -%}
65
+ {%- elif message.role == "assistant" and message.name is defined and message.name -%}
66
+ {%- set header.text = header.text ~ message.name -%}
67
+ {%- endif -%}
68
+ {%- set rtok = header.text -%}
69
+ {%- set rendered -%}
70
+
71
+ {%- if message.role == "assistant" and message.reasoning_content is defined and message.reasoning_content -%}
72
+ {{- rtok ~ "<|content_thinking|>" ~ message.reasoning_content ~ "<|end_message|>" -}}
73
+ {%- endif -%}
74
+
75
+ {%- if message.content is string and (message.content or not (message.role == "assistant" and (message.tool_calls or message.reasoning_content))) -%}
76
+ {{- rtok ~ "<|content_text|>" ~ message.content ~ "<|end_message|>" -}}
77
+ {%- elif message.content -%}
78
+ {%- for part in message.content -%}
79
+ {%- if part is string -%}
80
+ {{- rtok ~ "<|content_text|>" ~ part ~ "<|end_message|>" -}}
81
+ {%- elif part.type is not defined or part.type in ("text", "input_text") -%}
82
+ {%- set text_part = (part.text if part.text is defined and part.text is string else "") -%}
83
+ {{- rtok ~ "<|content_text|>" ~ text_part ~ "<|end_message|>" -}}
84
+ {%- elif part.type in ("image", "input_image", "image_url") -%}
85
+ {%- if message.role not in ("user", "tool") -%}
86
+ {{- raise_exception("Image content is only supported for user and tool messages") -}}
87
+ {%- endif -%}
88
+ {{- rtok ~ "<|content_image|><|unused_200054|><|end_message|>" -}}
89
+ {%- elif part.type in ("audio", "input_audio", "audio_url") -%}
90
+ {{- raise_exception("Inkling 2 does not accept audio input") -}}
91
+ {%- else -%}
92
+ {{- raise_exception("Unsupported content part type: " ~ part.type) -}}
93
+ {%- endif -%}
94
+ {%- endfor -%}
95
+ {%- endif -%}
96
+
97
+ {%- if message.role == "assistant" and message.tool_calls -%}
98
+ {%- for tc in message.tool_calls -%}
99
+ {%- set fn = tc.function -%}
100
+ {%- if fn.name is not defined or fn.name is not string -%}
101
+ {{- raise_exception("tool call function name must be a string") -}}
102
+ {%- endif -%}
103
+ {%- if message.name is defined and message.name and message.name != fn.name -%}
104
+ {{- raise_exception("assistant name " ~ message.name ~ " does not match its tool call " ~ fn.name) -}}
105
+ {%- endif -%}
106
+ {%- set args = fn.arguments if fn.arguments is defined and fn.arguments else {} -%}
107
+ {%- if args is string -%}
108
+ {{- raise_exception("tool call arguments must be a parsed object, not a JSON string; canonicalize upstream") -}}
109
+ {%- endif -%}
110
+ {%- if args is not mapping -%}
111
+ {{- raise_exception("tool call arguments must be an object") -}}
112
+ {%- endif -%}
113
+ {{- "<|message_model|>" ~ fn.name ~ "<|content_invoke_tool_json|>" -}}
114
+ {{- '{"name":' ~ (fn.name | tojson(sort_keys=true, separators=(",", ":"))) ~ ',"args":' -}}
115
+ {{- "{" -}}
116
+ {%- for key, value in args | dictsort(case_sensitive=true) -%}
117
+ {{- (key | tojson) ~ ":" ~ (value | tojson(separators=(",", ":"))) -}}
118
+ {%- if not loop.last -%},{%- endif -%}
119
+ {%- endfor -%}
120
+ {{- "}}<|end_message|>" -}}
121
+ {%- endfor -%}
122
+ {%- endif -%}
123
+
124
+ {%- if message.role == "assistant" -%}
125
+ {{- "<|content_model_end_sampling|>" -}}
126
+ {%- endif -%}
127
+ {%- endset -%}
128
+ {%- if message.role == "assistant" -%}
129
+ {%- generation -%}{{- rendered -}}{%- endgeneration -%}
130
+ {%- else -%}
131
+ {{- rendered -}}
132
+ {%- endif -%}
133
+ {%- endfor -%}
134
+
135
+ {%- if not state.effort_emitted -%}
136
+ {{- emit_thinking_effort() -}}
137
+ {%- endif -%}
config.json ADDED
@@ -0,0 +1,152 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Inkling2LoopForConditionalGeneration"
4
+ ],
5
+ "dtype": "bfloat16",
6
+ "eos_token_id": 200006,
7
+ "image_token_id": 200054,
8
+ "model_type": "inkling2_loop",
9
+ "mtp_config": {
10
+ "block_size": 8,
11
+ "head_dim": 128,
12
+ "hidden_size": 2304,
13
+ "intermediate_size": 9216,
14
+ "local_layer_ids": [
15
+ 0,
16
+ 2
17
+ ],
18
+ "markov_head_type": "rnn_hidden",
19
+ "markov_rank": 256,
20
+ "max_position_embeddings": 1048576,
21
+ "num_attention_heads": 24,
22
+ "num_hidden_layers": 3,
23
+ "num_key_value_heads": 8,
24
+ "o_bias": false,
25
+ "q_bias": false,
26
+ "rms_norm_eps": 1e-06,
27
+ "rope_theta": 1000000.0,
28
+ "sliding_window_size": 1024,
29
+ "target_layer_ids": [
30
+ 5,
31
+ 23,
32
+ 41,
33
+ 48
34
+ ],
35
+ "torch_dtype": "bfloat16"
36
+ },
37
+ "text_config": {
38
+ "attention_dropout": 0.0,
39
+ "attn_res_normalize_values": true,
40
+ "bos_token_id": null,
41
+ "conv_kernel_size": 4,
42
+ "d_rel": 4,
43
+ "dtype": "bfloat16",
44
+ "eos_token_id": 200006,
45
+ "final_logit_softcapping": 30.0,
46
+ "head_dim": 8,
47
+ "hidden_act": "silu",
48
+ "hidden_size": 32,
49
+ "initializer_range": 0.02,
50
+ "layer_types": [
51
+ "hybrid_sliding",
52
+ "hybrid",
53
+ "hybrid_sliding",
54
+ "hybrid",
55
+ "hybrid_sliding",
56
+ "hybrid",
57
+ "hybrid_sliding",
58
+ "hybrid"
59
+ ],
60
+ "local_layer_ids": [
61
+ 0,
62
+ 1,
63
+ 2,
64
+ 3,
65
+ 4,
66
+ 6,
67
+ 7,
68
+ 8,
69
+ 9,
70
+ 10,
71
+ 12,
72
+ 13,
73
+ 14,
74
+ 15,
75
+ 16,
76
+ 18,
77
+ 19,
78
+ 20,
79
+ 21,
80
+ 22
81
+ ],
82
+ "log_scaling_alpha": 0.1,
83
+ "log_scaling_n_floor": 8192,
84
+ "loop_config": {
85
+ "core_end": 6,
86
+ "core_start": 2,
87
+ "gate_rank": 8,
88
+ "is_varloop": false,
89
+ "num_passes": 3
90
+ },
91
+ "loop_gate_rank": 8,
92
+ "loop_layer_index_end": 6,
93
+ "loop_layer_index_start": 2,
94
+ "loop_num_passes": 3,
95
+ "max_position_embeddings": 1048576,
96
+ "mlp_gate_softcap": 4.0,
97
+ "mlp_up_softcap": 25.0,
98
+ "model_type": "inkling2_loop_text",
99
+ "moe_intermediate_size": 32,
100
+ "n_layers_per_attn_res_block": 2,
101
+ "n_routed_experts": 8,
102
+ "n_shared_experts": 0,
103
+ "num_attention_heads": 4,
104
+ "num_experts_per_tok": 2,
105
+ "num_hidden_layers": 8,
106
+ "num_key_value_heads": 2,
107
+ "o_bias": false,
108
+ "pad_token_id": null,
109
+ "q_bias": false,
110
+ "qk_l2_clip_threshold": 22.627416997969522,
111
+ "rel_extent": 32,
112
+ "rel_l2_clip_threshold": 8.0,
113
+ "rms_norm_eps": 1e-06,
114
+ "route_scale": 8.0,
115
+ "scoring_func": "sigmoid",
116
+ "sliding_window_size": 16,
117
+ "swa_head_dim": 8,
118
+ "swa_num_attention_heads": 4,
119
+ "swa_num_key_value_heads": 2,
120
+ "unpadded_vocab_size": 200058,
121
+ "use_cache": true,
122
+ "use_gate_bias": true,
123
+ "use_value_norm_no_gamma": true,
124
+ "vocab_size": 201024
125
+ },
126
+ "transformers_version": "5.24.0.dev0",
127
+ "vision_config": {
128
+ "attention_dropout": 0.0,
129
+ "hidden_act": "gelu",
130
+ "hidden_size": 32,
131
+ "initializer_range": 0.02,
132
+ "intermediate_size": 4096,
133
+ "layer_norm_eps": 1e-05,
134
+ "layer_scale_init_value": 0.1,
135
+ "mlp_ratio": 4.0,
136
+ "model_type": "inkling2_vision",
137
+ "num_attention_heads": 2,
138
+ "num_channels": 3,
139
+ "num_hidden_layers": 2,
140
+ "patch_size": 20,
141
+ "posemb_grid_size": 32,
142
+ "projection_hidden_dim": 64,
143
+ "projector_bias": false,
144
+ "rope_parameters": {
145
+ "rope_theta": 10000.0,
146
+ "rope_type": "axial"
147
+ },
148
+ "spatial_merge_size": 2,
149
+ "temporal_patch_size": 2,
150
+ "text_hidden_size": 32
151
+ }
152
+ }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_sample": true,
3
+ "eos_token_id": 200006,
4
+ "pad_token_id": 199999,
5
+ "temperature": 1.0,
6
+ "transformers_version": "5.24.0.dev0"
7
+ }
hf_quant_config.json ADDED
@@ -0,0 +1,293 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": {
3
+ "parameter_groups": [
4
+ {
5
+ "compute_dtype": "float32",
6
+ "output_dtype": "float32",
7
+ "required_activation_quantization": null,
8
+ "tensors": [
9
+ "model.llm.attn_res.attn_queries.0",
10
+ "model.llm.attn_res.attn_queries.1",
11
+ "model.llm.attn_res.attn_queries.2",
12
+ "model.llm.attn_res.attn_queries.3",
13
+ "model.llm.attn_res.attn_queries.4",
14
+ "model.llm.attn_res.attn_queries.5",
15
+ "model.llm.attn_res.attn_queries.6",
16
+ "model.llm.attn_res.mlp_queries.0",
17
+ "model.llm.attn_res.mlp_queries.1",
18
+ "model.llm.attn_res.mlp_queries.2",
19
+ "model.llm.attn_res.mlp_queries.3",
20
+ "model.llm.attn_res.mlp_queries.4",
21
+ "model.llm.attn_res.mlp_queries.5",
22
+ "model.llm.attn_res.mlp_queries.6",
23
+ "model.llm.attn_res.mlp_queries.7",
24
+ "model.llm.attn_res.output_query",
25
+ "model.llm.layers.0.mlp.gate.bias",
26
+ "model.llm.layers.1.mlp.gate.bias",
27
+ "model.llm.layers.2.mlp.gate.bias",
28
+ "model.llm.layers.3.mlp.gate.bias",
29
+ "model.llm.layers.4.mlp.gate.bias",
30
+ "model.llm.layers.5.mlp.gate.bias",
31
+ "model.llm.layers.6.mlp.gate.bias",
32
+ "model.llm.layers.7.mlp.gate.bias",
33
+ "model.llm.middle_cycle_cross_pass_kv_queries",
34
+ "model.llm.middle_cycle_cross_pass_queries"
35
+ ],
36
+ "weight_block_shape": [],
37
+ "weight_transform": "cast"
38
+ },
39
+ {
40
+ "compute_dtype": "bfloat16",
41
+ "output_dtype": "bfloat16",
42
+ "required_activation_quantization": null,
43
+ "tensors": [
44
+ "model.llm.embed.weight",
45
+ "model.llm.embed_norm.weight",
46
+ "model.llm.layers.0.attn.attn_sink_proj.weight",
47
+ "model.llm.layers.0.attn.gate_attn.weight",
48
+ "model.llm.layers.0.attn.k_sconv.weight",
49
+ "model.llm.layers.0.attn.q_l2_clip.scale",
50
+ "model.llm.layers.0.attn.rel_l2_clip.scale",
51
+ "model.llm.layers.0.attn.rel_logits_proj.proj",
52
+ "model.llm.layers.0.attn.v_sconv.weight",
53
+ "model.llm.layers.0.attn.wk_dv.weight",
54
+ "model.llm.layers.0.attn.wo_ud.weight",
55
+ "model.llm.layers.0.attn.wq_du.weight",
56
+ "model.llm.layers.0.attn.wr_du.weight",
57
+ "model.llm.layers.0.attn.wv_dv.weight",
58
+ "model.llm.layers.0.attn_norm.weight",
59
+ "model.llm.layers.0.attn_sconv.weight",
60
+ "model.llm.layers.0.mlp.gate.weight",
61
+ "model.llm.layers.0.mlp_norm.weight",
62
+ "model.llm.layers.0.mlp_sconv.weight",
63
+ "model.llm.layers.1.attn.attn_sink_proj.weight",
64
+ "model.llm.layers.1.attn.gate_attn.weight",
65
+ "model.llm.layers.1.attn.q_l2_clip.scale",
66
+ "model.llm.layers.1.attn.rel_l2_clip.scale",
67
+ "model.llm.layers.1.attn.rel_logits_proj.proj",
68
+ "model.llm.layers.1.attn.wk_dv.weight",
69
+ "model.llm.layers.1.attn.wo_ud.weight",
70
+ "model.llm.layers.1.attn.wq_du.weight",
71
+ "model.llm.layers.1.attn.wr_du.weight",
72
+ "model.llm.layers.1.attn.wv_dv.weight",
73
+ "model.llm.layers.1.attn_sconv.weight",
74
+ "model.llm.layers.1.mlp.gate.weight",
75
+ "model.llm.layers.1.mlp_norm.weight",
76
+ "model.llm.layers.1.mlp_sconv.weight",
77
+ "model.llm.layers.2.attn.attn_sink_proj.weight",
78
+ "model.llm.layers.2.attn.gate_attn.weight",
79
+ "model.llm.layers.2.attn.k_sconv.weight",
80
+ "model.llm.layers.2.attn.q_l2_clip.scale",
81
+ "model.llm.layers.2.attn.rel_l2_clip.scale",
82
+ "model.llm.layers.2.attn.rel_logits_proj.proj",
83
+ "model.llm.layers.2.attn.v_sconv.weight",
84
+ "model.llm.layers.2.attn.wk_dv.weight",
85
+ "model.llm.layers.2.attn.wo_ud.weight",
86
+ "model.llm.layers.2.attn.wq_du.weight",
87
+ "model.llm.layers.2.attn.wr_du.weight",
88
+ "model.llm.layers.2.attn.wv_dv.weight",
89
+ "model.llm.layers.2.attn_norm.weight",
90
+ "model.llm.layers.2.attn_sconv.weight",
91
+ "model.llm.layers.2.mlp.gate.weight",
92
+ "model.llm.layers.2.mlp_norm.weight",
93
+ "model.llm.layers.2.mlp_sconv.weight",
94
+ "model.llm.layers.3.attn.attn_sink_proj.weight",
95
+ "model.llm.layers.3.attn.gate_attn.weight",
96
+ "model.llm.layers.3.attn.q_l2_clip.scale",
97
+ "model.llm.layers.3.attn.rel_l2_clip.scale",
98
+ "model.llm.layers.3.attn.rel_logits_proj.proj",
99
+ "model.llm.layers.3.attn.wk_dv.weight",
100
+ "model.llm.layers.3.attn.wo_ud.weight",
101
+ "model.llm.layers.3.attn.wq_du.weight",
102
+ "model.llm.layers.3.attn.wr_du.weight",
103
+ "model.llm.layers.3.attn.wv_dv.weight",
104
+ "model.llm.layers.3.attn_sconv.weight",
105
+ "model.llm.layers.3.mlp.gate.weight",
106
+ "model.llm.layers.3.mlp_norm.weight",
107
+ "model.llm.layers.3.mlp_sconv.weight",
108
+ "model.llm.layers.4.attn.attn_sink_proj.weight",
109
+ "model.llm.layers.4.attn.gate_attn.weight",
110
+ "model.llm.layers.4.attn.k_sconv.weight",
111
+ "model.llm.layers.4.attn.q_l2_clip.scale",
112
+ "model.llm.layers.4.attn.rel_l2_clip.scale",
113
+ "model.llm.layers.4.attn.rel_logits_proj.proj",
114
+ "model.llm.layers.4.attn.v_sconv.weight",
115
+ "model.llm.layers.4.attn.wk_dv.weight",
116
+ "model.llm.layers.4.attn.wo_ud.weight",
117
+ "model.llm.layers.4.attn.wq_du.weight",
118
+ "model.llm.layers.4.attn.wr_du.weight",
119
+ "model.llm.layers.4.attn.wv_dv.weight",
120
+ "model.llm.layers.4.attn_norm.weight",
121
+ "model.llm.layers.4.attn_sconv.weight",
122
+ "model.llm.layers.4.mlp.gate.weight",
123
+ "model.llm.layers.4.mlp_norm.weight",
124
+ "model.llm.layers.4.mlp_sconv.weight",
125
+ "model.llm.layers.5.attn.attn_sink_proj.weight",
126
+ "model.llm.layers.5.attn.gate_attn.weight",
127
+ "model.llm.layers.5.attn.k_sconv.static_w",
128
+ "model.llm.layers.5.attn.k_sconv.weight_proj",
129
+ "model.llm.layers.5.attn.q_l2_clip.scale",
130
+ "model.llm.layers.5.attn.rel_l2_clip.scale",
131
+ "model.llm.layers.5.attn.rel_logits_proj.proj",
132
+ "model.llm.layers.5.attn.v_sconv.static_w",
133
+ "model.llm.layers.5.attn.v_sconv.weight_proj",
134
+ "model.llm.layers.5.attn.wk_dv.weight",
135
+ "model.llm.layers.5.attn.wo_ud.weight",
136
+ "model.llm.layers.5.attn.wq_du.weight",
137
+ "model.llm.layers.5.attn.wr_du.weight",
138
+ "model.llm.layers.5.attn.wv_dv.weight",
139
+ "model.llm.layers.5.attn_sconv.weight",
140
+ "model.llm.layers.5.k_input_norm.weight",
141
+ "model.llm.layers.5.mlp.gate.weight",
142
+ "model.llm.layers.5.mlp_norm.weight",
143
+ "model.llm.layers.5.mlp_sconv.weight",
144
+ "model.llm.layers.5.q_input_norm.weight",
145
+ "model.llm.layers.5.v_input_norm.weight",
146
+ "model.llm.layers.6.attn.attn_sink_proj.weight",
147
+ "model.llm.layers.6.attn.gate_attn.weight",
148
+ "model.llm.layers.6.attn.k_sconv.weight",
149
+ "model.llm.layers.6.attn.q_l2_clip.scale",
150
+ "model.llm.layers.6.attn.rel_l2_clip.scale",
151
+ "model.llm.layers.6.attn.rel_logits_proj.proj",
152
+ "model.llm.layers.6.attn.v_sconv.weight",
153
+ "model.llm.layers.6.attn.wk_dv.weight",
154
+ "model.llm.layers.6.attn.wo_ud.weight",
155
+ "model.llm.layers.6.attn.wq_du.weight",
156
+ "model.llm.layers.6.attn.wr_du.weight",
157
+ "model.llm.layers.6.attn.wv_dv.weight",
158
+ "model.llm.layers.6.attn_norm.weight",
159
+ "model.llm.layers.6.attn_sconv.weight",
160
+ "model.llm.layers.6.mlp.gate.weight",
161
+ "model.llm.layers.6.mlp_norm.weight",
162
+ "model.llm.layers.6.mlp_sconv.weight",
163
+ "model.llm.layers.7.attn.attn_sink_proj.weight",
164
+ "model.llm.layers.7.attn.gate_attn.weight",
165
+ "model.llm.layers.7.attn.q_l2_clip.scale",
166
+ "model.llm.layers.7.attn.rel_l2_clip.scale",
167
+ "model.llm.layers.7.attn.rel_logits_proj.proj",
168
+ "model.llm.layers.7.attn.wk_dv.weight",
169
+ "model.llm.layers.7.attn.wo_ud.weight",
170
+ "model.llm.layers.7.attn.wq_du.weight",
171
+ "model.llm.layers.7.attn.wr_du.weight",
172
+ "model.llm.layers.7.attn.wv_dv.weight",
173
+ "model.llm.layers.7.attn_sconv.weight",
174
+ "model.llm.layers.7.mlp.gate.weight",
175
+ "model.llm.layers.7.mlp_norm.weight",
176
+ "model.llm.layers.7.mlp_sconv.weight",
177
+ "model.llm.norm.weight",
178
+ "model.visual.backbone.embeddings.class_embedding",
179
+ "model.visual.backbone.embeddings.cls_position_embedding",
180
+ "model.visual.backbone.embeddings.patch_embedding.weight",
181
+ "model.visual.backbone.embeddings.position_embedding",
182
+ "model.visual.backbone.layers.0.ln_1.bias",
183
+ "model.visual.backbone.layers.0.ln_1.weight",
184
+ "model.visual.backbone.layers.0.ln_2.bias",
185
+ "model.visual.backbone.layers.0.ln_2.weight",
186
+ "model.visual.backbone.layers.0.ls_1.gamma",
187
+ "model.visual.backbone.layers.0.ls_2.gamma",
188
+ "model.visual.backbone.layers.0.mlp.c_fc.bias",
189
+ "model.visual.backbone.layers.0.mlp.c_fc.weight",
190
+ "model.visual.backbone.layers.0.mlp.c_proj.bias",
191
+ "model.visual.backbone.layers.0.mlp.c_proj.weight",
192
+ "model.visual.backbone.layers.0.self_attn.k_proj.bias",
193
+ "model.visual.backbone.layers.0.self_attn.k_proj.weight",
194
+ "model.visual.backbone.layers.0.self_attn.out_proj.bias",
195
+ "model.visual.backbone.layers.0.self_attn.out_proj.weight",
196
+ "model.visual.backbone.layers.0.self_attn.q_proj.bias",
197
+ "model.visual.backbone.layers.0.self_attn.q_proj.weight",
198
+ "model.visual.backbone.layers.0.self_attn.v_proj.bias",
199
+ "model.visual.backbone.layers.0.self_attn.v_proj.weight",
200
+ "model.visual.backbone.layers.1.ln_1.bias",
201
+ "model.visual.backbone.layers.1.ln_1.weight",
202
+ "model.visual.backbone.layers.1.ln_2.bias",
203
+ "model.visual.backbone.layers.1.ln_2.weight",
204
+ "model.visual.backbone.layers.1.ls_1.gamma",
205
+ "model.visual.backbone.layers.1.ls_2.gamma",
206
+ "model.visual.backbone.layers.1.mlp.c_fc.bias",
207
+ "model.visual.backbone.layers.1.mlp.c_fc.weight",
208
+ "model.visual.backbone.layers.1.mlp.c_proj.bias",
209
+ "model.visual.backbone.layers.1.mlp.c_proj.weight",
210
+ "model.visual.backbone.layers.1.self_attn.k_proj.bias",
211
+ "model.visual.backbone.layers.1.self_attn.k_proj.weight",
212
+ "model.visual.backbone.layers.1.self_attn.out_proj.bias",
213
+ "model.visual.backbone.layers.1.self_attn.out_proj.weight",
214
+ "model.visual.backbone.layers.1.self_attn.q_proj.bias",
215
+ "model.visual.backbone.layers.1.self_attn.q_proj.weight",
216
+ "model.visual.backbone.layers.1.self_attn.v_proj.bias",
217
+ "model.visual.backbone.layers.1.self_attn.v_proj.weight",
218
+ "model.visual.backbone.ln_pre.bias",
219
+ "model.visual.backbone.ln_pre.weight",
220
+ "model.visual.proj_in.weight",
221
+ "model.visual.proj_out.weight"
222
+ ],
223
+ "weight_block_shape": [],
224
+ "weight_transform": "cast"
225
+ },
226
+ {
227
+ "compute_dtype": "nvfp4",
228
+ "output_dtype": "uint8",
229
+ "required_activation_quantization": {
230
+ "amax_type": "per_token",
231
+ "block_size": 16,
232
+ "format": "nvfp4_e2m1",
233
+ "rounding": "nearest",
234
+ "scale_dtype": "float8_e4m3fn",
235
+ "scale_selection_rule": "fouroversix_mae"
236
+ },
237
+ "tensors": [
238
+ "model.llm.layers.0.mlp.experts.w13_weight",
239
+ "model.llm.layers.0.mlp.experts.w2_weight",
240
+ "model.llm.layers.1.mlp.experts.w13_weight",
241
+ "model.llm.layers.1.mlp.experts.w2_weight",
242
+ "model.llm.layers.2.mlp.experts.w13_weight",
243
+ "model.llm.layers.2.mlp.experts.w2_weight",
244
+ "model.llm.layers.3.mlp.experts.w13_weight",
245
+ "model.llm.layers.3.mlp.experts.w2_weight",
246
+ "model.llm.layers.4.mlp.experts.w13_weight",
247
+ "model.llm.layers.4.mlp.experts.w2_weight",
248
+ "model.llm.layers.5.mlp.experts.w13_weight",
249
+ "model.llm.layers.5.mlp.experts.w2_weight"
250
+ ],
251
+ "weight_block_shape": [
252
+ 1,
253
+ 16
254
+ ],
255
+ "weight_scale_format": "amax_denominator",
256
+ "weight_transform": "nvfp4_nearest"
257
+ },
258
+ {
259
+ "compute_dtype": "mxfp8",
260
+ "output_dtype": "float8_e4m3fn",
261
+ "required_activation_quantization": {
262
+ "block_size": 32,
263
+ "format": "mxfp8_e4m3",
264
+ "rounding": "nearest",
265
+ "scale_dtype": "ue8m0"
266
+ },
267
+ "tensors": [
268
+ "model.llm.layers.6.mlp.experts.w13_weight",
269
+ "model.llm.layers.6.mlp.experts.w2_weight",
270
+ "model.llm.layers.7.mlp.experts.w13_weight",
271
+ "model.llm.layers.7.mlp.experts.w2_weight"
272
+ ],
273
+ "weight_block_shape": [
274
+ 32,
275
+ 32
276
+ ],
277
+ "weight_transform": "mxfp8_nearest"
278
+ },
279
+ {
280
+ "compute_dtype": "bfloat16",
281
+ "fwd_output_dtype": "float32",
282
+ "output_dtype": "bfloat16",
283
+ "required_activation_quantization": null,
284
+ "tensors": [
285
+ "model.llm.unembed.weight"
286
+ ],
287
+ "weight_block_shape": [],
288
+ "weight_transform": "cast"
289
+ }
290
+ ]
291
+ },
292
+ "quant_method": "inkling2"
293
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3376ee624f277a4222bc0beeeedc62f2b6eba0f3d83a96dc387f3e693e42825c
3
+ size 27279560
processor_config.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "image_processor": {
3
+ "do_convert_rgb": true,
4
+ "do_normalize": true,
5
+ "do_rescale": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "Inkling2ImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 40,
20
+ "width": 40
21
+ }
22
+ },
23
+ "image_token": "<|unused_200054|>",
24
+ "processor_class": "Inkling2Processor"
25
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1bc3492b4ccc4ac57e227c13c259c36851d159a592bdabb101e5fd278ac9f600
3
+ size 26178846
tokenizer_config.json ADDED
@@ -0,0 +1,85 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "begin_of_text": "<|begin_of_text|>",
4
+ "clean_up_tokenization_spaces": false,
5
+ "content_image": "<|content_image|>",
6
+ "content_invoke_tool_json": "<|content_invoke_tool_json|>",
7
+ "content_invoke_tool_text": "<|content_invoke_tool_text|>",
8
+ "content_model_end_sampling": "<|content_model_end_sampling|>",
9
+ "content_text": "<|content_text|>",
10
+ "content_thinking": "<|content_thinking|>",
11
+ "content_tool_error": "<|content_tool_error|>",
12
+ "content_xml": "<|content_xml|>",
13
+ "end_message": "<|end_message|>",
14
+ "endoftext": "<|endoftext|>",
15
+ "eos_token": "<|content_model_end_sampling|>",
16
+ "fix_mistral_regex": false,
17
+ "is_local": false,
18
+ "local_files_only": false,
19
+ "message_model": "<|message_model|>",
20
+ "message_system": "<|message_system|>",
21
+ "message_tool": "<|message_tool|>",
22
+ "message_user": "<|message_user|>",
23
+ "model_max_length": 1048576,
24
+ "model_specific_special_tokens": {
25
+ "begin_of_text": "<|begin_of_text|>",
26
+ "content_image": "<|content_image|>",
27
+ "content_invoke_tool_json": "<|content_invoke_tool_json|>",
28
+ "content_invoke_tool_text": "<|content_invoke_tool_text|>",
29
+ "content_model_end_sampling": "<|content_model_end_sampling|>",
30
+ "content_text": "<|content_text|>",
31
+ "content_thinking": "<|content_thinking|>",
32
+ "content_tool_error": "<|content_tool_error|>",
33
+ "content_xml": "<|content_xml|>",
34
+ "end_message": "<|end_message|>",
35
+ "endoftext": "<|endoftext|>",
36
+ "message_model": "<|message_model|>",
37
+ "message_system": "<|message_system|>",
38
+ "message_tool": "<|message_tool|>",
39
+ "message_user": "<|message_user|>"
40
+ },
41
+ "pad_token": "<|endoftext|>",
42
+ "processor_class": "Inkling2Processor",
43
+ "response_template": {
44
+ "defaults": {
45
+ "role": "assistant"
46
+ },
47
+ "fields": {
48
+ "content": {
49
+ "close": "<|end_message|>",
50
+ "content": "text",
51
+ "content_args": {
52
+ "strip": false
53
+ },
54
+ "join": "",
55
+ "open_pattern": "(?:<\\|message_model\\|>[^<]*)?<\\|content_(?:text|tool_error|invoke_tool_text)\\|>",
56
+ "repeats": true
57
+ },
58
+ "reasoning_content": {
59
+ "close": "<|end_message|>",
60
+ "content": "text",
61
+ "content_args": {
62
+ "strip": false
63
+ },
64
+ "join": "",
65
+ "open_pattern": "(?:<\\|message_model\\|>[^<]*)?<\\|content_thinking\\|>",
66
+ "repeats": true
67
+ },
68
+ "tool_calls": {
69
+ "close": "<|end_message|>",
70
+ "content": "json",
71
+ "open_pattern": "(?:<\\|message_model\\|>[^<]*)?<\\|content_invoke_tool_json\\|>",
72
+ "repeats": true,
73
+ "transform": {
74
+ "function": {
75
+ "arguments": "{content.args}",
76
+ "name": "{content.name}"
77
+ },
78
+ "type": "function"
79
+ }
80
+ }
81
+ },
82
+ "start_anchor_pattern": "<\\|end_message\\|>|<\\|content_model_end_sampling\\|>"
83
+ },
84
+ "tokenizer_class": "TokenizersBackend"
85
+ }