devankit7873 commited on
Commit
f622acc
·
verified ·
1 Parent(s): f61484d

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
LICENSE ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ Apache License
3
+ Version 2.0, January 2004
4
+ http://www.apache.org/licenses/
5
+
6
+ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
7
+
8
+ 1. Definitions.
9
+
10
+ "License" shall mean the terms and conditions for use, reproduction,
11
+ and distribution as defined by Sections 1 through 9 of this document.
12
+
13
+ "Licensor" shall mean the copyright owner or entity authorized by
14
+ the copyright owner that is granting the License.
15
+
16
+ "Legal Entity" shall mean the union of the acting entity and all
17
+ other entities that control, are controlled by, or are under common
18
+ control with that entity. For the purposes of this definition,
19
+ "control" means (i) the power, direct or indirect, to cause the
20
+ direction or management of such entity, whether by contract or
21
+ otherwise, or (ii) ownership of fifty percent (50%) or more of the
22
+ outstanding shares, or (iii) beneficial ownership of such entity.
23
+
24
+ "You" (or "Your") shall mean an individual or Legal Entity
25
+ exercising permissions granted by this License.
26
+
27
+ "Source" form shall mean the preferred form for making modifications,
28
+ including but not limited to software source code, documentation
29
+ source, and configuration files.
30
+
31
+ "Object" form shall mean any form resulting from mechanical
32
+ transformation or translation of a Source form, including but
33
+ not limited to compiled object code, generated documentation,
34
+ and conversions to other media types.
35
+
36
+ "Work" shall mean the work of authorship, whether in Source or
37
+ Object form, made available under the License, as indicated by a
38
+ copyright notice that is included in or attached to the work
39
+ (an example is provided in the Appendix below).
40
+
41
+ "Derivative Works" shall mean any work, whether in Source or Object
42
+ form, that is based on (or derived from) the Work and for which the
43
+ editorial revisions, annotations, elaborations, or other modifications
44
+ represent, as a whole, an original work of authorship. For the purposes
45
+ of this License, Derivative Works shall not include works that remain
46
+ separable from, or merely link (or bind by name) to the interfaces of,
47
+ the Work and Derivative Works thereof.
48
+
49
+ "Contribution" shall mean any work of authorship, including
50
+ the original version of the Work and any modifications or additions
51
+ to that Work or Derivative Works thereof, that is intentionally
52
+ submitted to Licensor for inclusion in the Work by the copyright owner
53
+ or by an individual or Legal Entity authorized to submit on behalf of
54
+ the copyright owner. For the purposes of this definition, "submitted"
55
+ means any form of electronic, verbal, or written communication sent
56
+ to the Licensor or its representatives, including but not limited to
57
+ communication on electronic mailing lists, source code control systems,
58
+ and issue tracking systems that are managed by, or on behalf of, the
59
+ Licensor for the purpose of discussing and improving the Work, but
60
+ excluding communication that is conspicuously marked or otherwise
61
+ designated in writing by the copyright owner as "Not a Contribution."
62
+
63
+ "Contributor" shall mean Licensor and any individual or Legal Entity
64
+ on behalf of whom a Contribution has been received by Licensor and
65
+ subsequently incorporated within the Work.
66
+
67
+ 2. Grant of Copyright License. Subject to the terms and conditions of
68
+ this License, each Contributor hereby grants to You a perpetual,
69
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
70
+ copyright license to reproduce, prepare Derivative Works of,
71
+ publicly display, publicly perform, sublicense, and distribute the
72
+ Work and such Derivative Works in Source or Object form.
73
+
74
+ 3. Grant of Patent License. Subject to the terms and conditions of
75
+ this License, each Contributor hereby grants to You a perpetual,
76
+ worldwide, non-exclusive, no-charge, royalty-free, irrevocable
77
+ (except as stated in this section) patent license to make, have made,
78
+ use, offer to sell, sell, import, and otherwise transfer the Work,
79
+ where such license applies only to those patent claims licensable
80
+ by such Contributor that are necessarily infringed by their
81
+ Contribution(s) alone or by combination of their Contribution(s)
82
+ with the Work to which such Contribution(s) was submitted. If You
83
+ institute patent litigation against any entity (including a
84
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
85
+ or a Contribution incorporated within the Work constitutes direct
86
+ or contributory patent infringement, then any patent licenses
87
+ granted to You under this License for that Work shall terminate
88
+ as of the date such litigation is filed.
89
+
90
+ 4. Redistribution. You may reproduce and distribute copies of the
91
+ Work or Derivative Works thereof in any medium, with or without
92
+ modifications, and in Source or Object form, provided that You
93
+ meet the following conditions:
94
+
95
+ (a) You must give any other recipients of the Work or
96
+ Derivative Works a copy of this License; and
97
+
98
+ (b) You must cause any modified files to carry prominent notices
99
+ stating that You changed the files; and
100
+
101
+ (c) You must retain, in the Source form of any Derivative Works
102
+ that You distribute, all copyright, patent, trademark, and
103
+ attribution notices from the Source form of the Work,
104
+ excluding those notices that do not pertain to any part of
105
+ the Derivative Works; and
106
+
107
+ (d) If the Work includes a "NOTICE" text file as part of its
108
+ distribution, then any Derivative Works that You distribute must
109
+ include a readable copy of the attribution notices contained
110
+ within such NOTICE file, excluding those notices that do not
111
+ pertain to any part of the Derivative Works, in at least one
112
+ of the following places: within a NOTICE text file distributed
113
+ as part of the Derivative Works; within the Source form or
114
+ documentation, if provided along with the Derivative Works; or,
115
+ within a display generated by the Derivative Works, if and
116
+ wherever such third-party notices normally appear. The contents
117
+ of the NOTICE file are for informational purposes only and
118
+ do not modify the License. You may add Your own attribution
119
+ notices within Derivative Works that You distribute, alongside
120
+ or as an addendum to the NOTICE text from the Work, provided
121
+ that such additional attribution notices cannot be construed
122
+ as modifying the License.
123
+
124
+ You may add Your own copyright statement to Your modifications and
125
+ may provide additional or different license terms and conditions
126
+ for use, reproduction, or distribution of Your modifications, or
127
+ for any such Derivative Works as a whole, provided Your use,
128
+ reproduction, and distribution of the Work otherwise complies with
129
+ the conditions stated in this License.
130
+
131
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
132
+ any Contribution intentionally submitted for inclusion in the Work
133
+ by You to the Licensor shall be under the terms and conditions of
134
+ this License, without any additional terms or conditions.
135
+ Notwithstanding the above, nothing herein shall supersede or modify
136
+ the terms of any separate license agreement you may have executed
137
+ with Licensor regarding such Contributions.
138
+
139
+ 6. Trademarks. This License does not grant permission to use the trade
140
+ names, trademarks, service marks, or product names of the Licensor,
141
+ except as required for reasonable and customary use in describing the
142
+ origin of the Work and reproducing the content of the NOTICE file.
143
+
144
+ 7. Disclaimer of Warranty. Unless required by applicable law or
145
+ agreed to in writing, Licensor provides the Work (and each
146
+ Contributor provides its Contributions) on an "AS IS" BASIS,
147
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
148
+ implied, including, without limitation, any warranties or conditions
149
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
150
+ PARTICULAR PURPOSE. You are solely responsible for determining the
151
+ appropriateness of using or redistributing the Work and assume any
152
+ risks associated with Your exercise of permissions under this License.
153
+
154
+ 8. Limitation of Liability. In no event and under no legal theory,
155
+ whether in tort (including negligence), contract, or otherwise,
156
+ unless required by applicable law (such as deliberate and grossly
157
+ negligent acts) or agreed to in writing, shall any Contributor be
158
+ liable to You for damages, including any direct, indirect, special,
159
+ incidental, or consequential damages of any character arising as a
160
+ result of this License or out of the use or inability to use the
161
+ Work (including but not limited to damages for loss of goodwill,
162
+ work stoppage, computer failure or malfunction, or any and all
163
+ other commercial damages or losses), even if such Contributor
164
+ has been advised of the possibility of such damages.
165
+
166
+ 9. Accepting Warranty or Additional Liability. While redistributing
167
+ the Work or Derivative Works thereof, You may choose to offer,
168
+ and charge a fee for, acceptance of support, warranty, indemnity,
169
+ or other liability obligations and/or rights consistent with this
170
+ License. However, in accepting such obligations, You may act only
171
+ on Your own behalf and on Your sole responsibility, not on behalf
172
+ of any other Contributor, and only if You agree to indemnify,
173
+ defend, and hold each Contributor harmless for any liability
174
+ incurred by, or claims asserted against, such Contributor by reason
175
+ of your accepting any such warranty or additional liability.
176
+
177
+ END OF TERMS AND CONDITIONS
178
+
179
+ APPENDIX: How to apply the Apache License to your work.
180
+
181
+ To apply the Apache License to your work, attach the following
182
+ boilerplate notice, with the fields enclosed by brackets "[]"
183
+ replaced with your own identifying information. (Don't include
184
+ the brackets!) The text should be enclosed in the appropriate
185
+ comment syntax for the file format. We also recommend that a
186
+ file or class name and description of purpose be included on the
187
+ same "printed page" as the copyright notice for easier
188
+ identification within third-party archives.
189
+
190
+ Copyright 2024 Alibaba Cloud
191
+
192
+ Licensed under the Apache License, Version 2.0 (the "License");
193
+ you may not use this file except in compliance with the License.
194
+ You may obtain a copy of the License at
195
+
196
+ http://www.apache.org/licenses/LICENSE-2.0
197
+
198
+ Unless required by applicable law or agreed to in writing, software
199
+ distributed under the License is distributed on an "AS IS" BASIS,
200
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
201
+ See the License for the specific language governing permissions and
202
+ limitations under the License.
README.md ADDED
@@ -0,0 +1,142 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: Qwen/Qwen2.5-1.5B-Instruct
4
+ library_name: transformers
5
+ tags:
6
+ - transformers
7
+ - bitsandbytes
8
+ - 4-bit
9
+ - quantized
10
+ - pytorch
11
+ - qwen
12
+ - qwen2
13
+ - decision-making
14
+ - system-one
15
+ - reasoning
16
+ - jev
17
+ - jevbench
18
+ - joint-expected-value
19
+ - classification
20
+ - safetensors
21
+ - agent
22
+ - fast-inference
23
+ datasets:
24
+ - SargeDev/jev-distill-corpus-v3
25
+ metrics:
26
+ - accuracy
27
+ - tvd
28
+ - kl_divergence
29
+ pipeline_tag: text-classification
30
+ ---
31
+
32
+ <div align="center">
33
+
34
+ # ⚡ QJev-1.5B-4bit: 4-Bit NF4 Quantized Decision Model
35
+ ### High-Speed System 1 Decision Engine (1.15 GB Weights • 1.07 GB VRAM)
36
+
37
+ [![Hugging Face](https://img.shields.io/badge/Hugging%20Face-QJev--1.5B--4bit-FFD21E?style=for-the-badge&logo=huggingface&logoColor=black)](https://huggingface.co/devankit7873/QJev-1.5B-4bit)
38
+ [![GitHub](https://img.shields.io/badge/GitHub-QJev--1.5B-181717?style=for-the-badge&logo=github)](https://github.com/AnkitNayak-dev/QJev-1.5B)
39
+ [![JevBench](https://img.shields.io/badge/JevBench-71.5%25%20Score-8A2BE2?style=for-the-badge)](https://jevbench.github.io/)
40
+ [![License](https://img.shields.io/badge/License-Apache%202.0-2EA44F?style=for-the-badge)](LICENSE)
41
+
42
+ ---
43
+
44
+ ### 📦 Part of the QJev Model Family:
45
+ - **Flagship Standalone (Fused):** [devankit7873/QJev-1.5B](https://huggingface.co/devankit7873/QJev-1.5B) (3.08 GB)
46
+ - **Lightweight LoRA Adapter:** [devankit7873/QJev-1.5B-LoRA](https://huggingface.co/devankit7873/QJev-1.5B-LoRA) (20 MB)
47
+ - **4-Bit Quantized:** [devankit7873/QJev-1.5B-4bit](https://huggingface.co/devankit7873/QJev-1.5B-4bit) (1.15 GB, 1.07 GB VRAM)
48
+
49
+ ---
50
+
51
+ </div>
52
+
53
+ ## 💡 Overview
54
+
55
+ **`QJev-1.5B-4bit`** is the **ultra-efficient 4-bit NormalFloat (NF4) quantized release** of `QJev-1.5B`.
56
+
57
+ By compressing the 1.54B parameter transformer into 4-bit precision via `bitsandbytes`, the memory footprint drops from 3.4 GB down to **1.07 GB VRAM**, enabling real-time System 1 decisions on ultra-low-cost cloud instances, edge devices, and consumer laptops while preserving calibrated decision accuracy.
58
+
59
+ - **Storage Size:** **1.15 GB** (`model.safetensors`)
60
+ - **Active VRAM Footprint:** **1.07 GB**
61
+ - **Decision Speed:** ~35 ms forward pass
62
+ - **JevBench Metamorphic Accuracy:** Maintained within margin of full precision.
63
+
64
+ ---
65
+
66
+ ## 🚀 Quickstart: Direct 4-Bit Inference
67
+
68
+ ```bash
69
+ pip install torch transformers safetensors accelerate bitsandbytes
70
+ ```
71
+
72
+ ```python
73
+ import os
74
+ import torch
75
+ import torch.nn as nn
76
+ import torch.nn.functional as F
77
+ from transformers import AutoTokenizer, AutoModelForCausalLM
78
+ from safetensors.torch import load_file
79
+ from huggingface_hub import hf_hub_download
80
+
81
+ class DecisionHead(nn.Module):
82
+ def __init__(self, hidden_size: int = 1536, intermediate_size: int = 512, max_choices: int = 10, dropout: float = 0.1):
83
+ super().__init__()
84
+ self.max_choices = max_choices
85
+ self.norm = nn.LayerNorm(hidden_size)
86
+ self.dense = nn.Linear(hidden_size, intermediate_size)
87
+ self.act = nn.GELU()
88
+ self.dropout = nn.Dropout(dropout)
89
+ self.out_proj = nn.Linear(intermediate_size, max_choices)
90
+
91
+ def forward(self, x: torch.Tensor, num_choices: torch.Tensor = None) -> torch.Tensor:
92
+ h = self.dropout(self.act(self.dense(self.norm(x))))
93
+ logits = self.out_proj(h)
94
+ if num_choices is not None:
95
+ mask = torch.arange(self.max_choices, device=logits.device).unsqueeze(0) >= num_choices.unsqueeze(1)
96
+ logits = logits.masked_fill(mask, -1e4)
97
+ return logits
98
+
99
+ # 1. Load 4-Bit Model directly (Transformers handles bitsandbytes automatically)
100
+ model_id = "devankit7873/QJev-1.5B-4bit"
101
+ device = "cuda" if torch.cuda.is_available() else "cpu"
102
+
103
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
104
+ backbone = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto")
105
+
106
+ # 2. Attach Decision Head
107
+ head = DecisionHead().to(device=device, dtype=torch.bfloat16 if torch.cuda.is_available() else torch.float32)
108
+ head_path = hf_hub_download(repo_id=model_id, filename="decision_head.safetensors")
109
+ head.load_state_dict(load_file(head_path))
110
+ head.eval()
111
+
112
+ # 3. Predict with Calibrated Probabilities
113
+ context = "Production database primary replica CPU spiked to 100% and connection pool is exhausted."
114
+ choices = [
115
+ "Trigger immediate automated failover to standby replica",
116
+ "Restart the primary database daemon immediately",
117
+ "Double connection pool limits without scaling compute",
118
+ "Silence alerts and monitor for next 15 minutes"
119
+ ]
120
+
121
+ opts_str = "\n".join([f"[{i+1}] {c}" for i, c in enumerate(choices)])
122
+ prompt = (
123
+ "<|im_start|>system\nYou are an expert decision engine. Evaluate the context and select the optimal decision.<|im_end|>\n"
124
+ f"<|im_start|>user\nContext:\n{context}\n\nQuestion:\nSelect the most appropriate option:\n\nOptions:\n{opts_str}<|im_end|>\n"
125
+ "<|im_start|>assistant\nDecision:"
126
+ )
127
+
128
+ inputs = tokenizer(prompt, return_tensors="pt").to(device)
129
+ with torch.no_grad():
130
+ transformer = getattr(backbone, "model", backbone)
131
+ out = transformer(inputs["input_ids"], attention_mask=inputs["attention_mask"])
132
+ logits = head(out.last_hidden_state[:, -1, :], num_choices=torch.tensor([len(choices)], device=device))
133
+ probs = F.softmax(logits[0, :len(choices)], dim=-1).cpu().tolist()
134
+
135
+ for choice, prob in zip(choices, probs):
136
+ print(f"{choice}: {prob * 100:.2f}%")
137
+ ```
138
+
139
+ ---
140
+
141
+ ## 📄 License
142
+ Released under the **Apache 2.0 License**.
chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
config.json ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2ForCausalLM"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "dtype": "bfloat16",
8
+ "eos_token_id": 151645,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 1536,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 8960,
13
+ "layer_types": [
14
+ "full_attention",
15
+ "full_attention",
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention",
36
+ "full_attention",
37
+ "full_attention",
38
+ "full_attention",
39
+ "full_attention",
40
+ "full_attention",
41
+ "full_attention"
42
+ ],
43
+ "max_position_embeddings": 32768,
44
+ "max_window_layers": 21,
45
+ "model_type": "qwen2",
46
+ "num_attention_heads": 12,
47
+ "num_hidden_layers": 28,
48
+ "num_key_value_heads": 2,
49
+ "pad_token_id": null,
50
+ "quantization_config": {
51
+ "_load_in_4bit": true,
52
+ "_load_in_8bit": false,
53
+ "bnb_4bit_compute_dtype": "bfloat16",
54
+ "bnb_4bit_quant_storage": "uint8",
55
+ "bnb_4bit_quant_type": "nf4",
56
+ "bnb_4bit_use_double_quant": false,
57
+ "llm_int8_enable_fp32_cpu_offload": false,
58
+ "llm_int8_has_fp16_weight": false,
59
+ "llm_int8_skip_modules": null,
60
+ "llm_int8_threshold": 6.0,
61
+ "load_in_4bit": true,
62
+ "load_in_8bit": false,
63
+ "quant_method": "bitsandbytes"
64
+ },
65
+ "rms_norm_eps": 1e-06,
66
+ "rope_parameters": {
67
+ "rope_theta": 1000000.0,
68
+ "rope_type": "default"
69
+ },
70
+ "sliding_window": null,
71
+ "tie_word_embeddings": true,
72
+ "transformers_version": "5.22.0",
73
+ "use_cache": true,
74
+ "use_sliding_window": false,
75
+ "vocab_size": 151936
76
+ }
decision_config.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "qwen_decision_merged",
3
+ "base_model": "Qwen/Qwen2.5-1.5B-Instruct",
4
+ "is_lora_merged": true,
5
+ "hidden_size": 1536,
6
+ "head_intermediate_size": 512,
7
+ "max_choices": 10,
8
+ "dropout": 0.1,
9
+ "activation": "gelu",
10
+ "output_type": "jev_probabilities",
11
+ "training_corpus": "SargeDev/jev-distill-corpus-v3",
12
+ "training_samples": 655806,
13
+ "phase": 2,
14
+ "metrics": {
15
+ "test_split": {
16
+ "samples": 1500,
17
+ "eval_loss": 0.1033,
18
+ "top1_accuracy": 70.13,
19
+ "mean_kl_divergence": 0.1033,
20
+ "mean_tvd": 0.0921,
21
+ "latency_ms_per_sample": 31.49,
22
+ "total_eval_time_sec": 47.24
23
+ },
24
+ "ood_split": {
25
+ "samples": 1500,
26
+ "eval_loss": 0.6972,
27
+ "top1_accuracy": 80.07,
28
+ "mean_kl_divergence": 0.6972,
29
+ "mean_tvd": 0.2389,
30
+ "latency_ms_per_sample": 33.64,
31
+ "total_eval_time_sec": 50.46
32
+ }
33
+ }
34
+ }
decision_head.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f78e3ea84fe461810423d1de3e3e179ff0be8d7c40d739200781b939710254fc
3
+ size 3183276
generation_config.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.1,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "5.22.0"
14
+ }
handler.py ADDED
@@ -0,0 +1,181 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ import re
3
+ from typing import Dict, List, Any, Union
4
+ import torch
5
+ import torch.nn as nn
6
+ import torch.nn.functional as F
7
+ from transformers import AutoTokenizer, AutoModelForCausalLM
8
+ from safetensors.torch import load_file
9
+
10
+
11
+ class DecisionHead(nn.Module):
12
+ def __init__(self, hidden_size: int = 1536, intermediate_size: int = 512, max_choices: int = 10, dropout: float = 0.1):
13
+ super().__init__()
14
+ self.max_choices = max_choices
15
+ self.norm = nn.LayerNorm(hidden_size)
16
+ self.dense = nn.Linear(hidden_size, intermediate_size)
17
+ self.act = nn.GELU()
18
+ self.dropout = nn.Dropout(dropout)
19
+ self.out_proj = nn.Linear(intermediate_size, max_choices)
20
+ self.score_proj = nn.Linear(intermediate_size, 1)
21
+
22
+ def forward(self, x: torch.Tensor, num_choices: torch.Tensor = None) -> torch.Tensor:
23
+ h = self.norm(x)
24
+ h = self.act(self.dense(h))
25
+ h = self.dropout(h)
26
+ logits = self.out_proj(h)
27
+ if num_choices is not None:
28
+ batch_size = logits.size(0)
29
+ seq_range = torch.arange(self.max_choices, device=logits.device).unsqueeze(0).expand(batch_size, -1)
30
+ mask = seq_range >= num_choices.unsqueeze(1)
31
+ logits = logits.masked_fill(mask, -1e4)
32
+ return logits
33
+
34
+
35
+ class EndpointHandler:
36
+ """
37
+ Hugging Face Inference Endpoints Custom Handler for QJev-1.5B.
38
+ Accepts:
39
+ {
40
+ "inputs": "Situation or context description",
41
+ "parameters": {
42
+ "choices": ["Choice 1", "Choice 2", ...]
43
+ }
44
+ }
45
+ OR:
46
+ {
47
+ "context": "...",
48
+ "question": "...",
49
+ "choices": [...]
50
+ }
51
+ Returns standard text-classification scores:
52
+ [
53
+ {"label": "Choice 1", "score": 0.8672},
54
+ {"label": "Choice 2", "score": 0.0703},
55
+ ...
56
+ ]
57
+ """
58
+ def __init__(self, path: str = ""):
59
+ self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
60
+ dtype = torch.bfloat16 if torch.cuda.is_available() and torch.cuda.is_bf16_supported() else torch.float32
61
+
62
+ # 1. Load Tokenizer
63
+ self.tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True)
64
+ if self.tokenizer.pad_token is None:
65
+ self.tokenizer.pad_token = self.tokenizer.eos_token
66
+
67
+ # 2. Load Base Transformer Backbone
68
+ self.backbone = AutoModelForCausalLM.from_pretrained(
69
+ path,
70
+ torch_dtype=dtype,
71
+ device_map=self.device
72
+ )
73
+ self.backbone.eval()
74
+
75
+ # 3. Load MLP Decision Head
76
+ self.head = DecisionHead().to(device=self.device, dtype=dtype)
77
+ head_path = os.path.join(path, "decision_head.safetensors")
78
+ if not os.path.exists(head_path):
79
+ head_path = os.path.join(path, "decision_head.pt")
80
+
81
+ if os.path.exists(head_path):
82
+ if head_path.endswith(".safetensors"):
83
+ state = load_file(head_path)
84
+ else:
85
+ ckpt = torch.load(head_path, map_location="cpu")
86
+ state = ckpt.get("head_state_dict", ckpt)
87
+ self.head.load_state_dict(state)
88
+ self.head.eval()
89
+
90
+ def _extract_choices_from_text(self, text: str) -> List[str]:
91
+ # Try extracting patterns like "[1] ...", "1. ...", "Option A: ..."
92
+ lines = [line.strip() for line in text.split("\n") if line.strip()]
93
+ choices = []
94
+ for line in lines:
95
+ m = re.match(r"^(?:\[?\d+\]?[\.\)]?|[A-E][\.\)]|Option\s+[A-E]:?)\s*(.+)$", line, re.IGNORECASE)
96
+ if m:
97
+ choices.append(m.group(1).strip())
98
+ if len(choices) >= 2:
99
+ return choices[:10]
100
+ return ["Approve / Yes", "Reject / No"]
101
+
102
+ def _format_prompt(self, context: str, question: str, choices: List[str]) -> str:
103
+ choices_str = "\n".join([f"[{i+1}] {c}" for i, c in enumerate(choices)])
104
+ q_str = f"\n\nQuestion:\n{question.strip()}" if question else "\n\nQuestion:\nSelect the most appropriate option:"
105
+ return (
106
+ "<|im_start|>system\n"
107
+ "You are an expert decision judge. Evaluate the situation and determine the likelihood for each option.<|im_end|>\n"
108
+ f"<|im_start|>user\nContext:\n{context.strip()}{q_str}\n\nOptions:\n{choices_str}<|im_end|>\n"
109
+ "<|im_start|>assistant\nDecision:"
110
+ )
111
+
112
+ def _process_single(self, item: Dict[str, Any]) -> List[Dict[str, float]]:
113
+ # Parse inputs
114
+ context = ""
115
+ question = ""
116
+ choices = []
117
+
118
+ if isinstance(item, str):
119
+ context = item
120
+ choices = self._extract_choices_from_text(item)
121
+ elif isinstance(item, dict):
122
+ # Check standard Hugging Face inputs/parameters format
123
+ if "inputs" in item:
124
+ context = str(item.get("inputs", ""))
125
+ params = item.get("parameters", {}) or {}
126
+ if isinstance(params, dict):
127
+ choices = params.get("choices", [])
128
+ question = params.get("question", "")
129
+
130
+ # Check alternative keys
131
+ if not context and "context" in item:
132
+ context = str(item.get("context", ""))
133
+ if not question and "question" in item:
134
+ question = str(item.get("question", ""))
135
+ if not choices and "choices" in item:
136
+ choices = item.get("choices", [])
137
+ if not choices and "options" in item:
138
+ choices = item.get("options", [])
139
+
140
+ # Fallback if no choices provided
141
+ if not choices:
142
+ choices = self._extract_choices_from_text(context)
143
+
144
+ # Enforce max 10 choices (architecture limit)
145
+ choices = [str(c) for c in choices][:10]
146
+ if len(choices) < 2:
147
+ choices = ["Approve / Yes", "Reject / No"]
148
+
149
+ num_c = len(choices)
150
+ prompt = self._format_prompt(context, question, choices)
151
+ inputs = self.tokenizer(prompt, return_tensors="pt").to(self.device)
152
+
153
+ with torch.no_grad():
154
+ transformer = getattr(self.backbone, "model", self.backbone)
155
+ out = transformer(inputs["input_ids"], attention_mask=inputs["attention_mask"])
156
+ last_hidden = out.last_hidden_state[:, -1, :]
157
+ logits = self.head(last_hidden, num_choices=torch.tensor([num_c], device=self.device))
158
+ probs = F.softmax(logits[0, :num_c], dim=-1).cpu().tolist()
159
+
160
+ return [{"label": choice, "score": round(prob, 4)} for choice, prob in zip(choices, probs)]
161
+
162
+ def __call__(self, data: Union[Dict[str, Any], List[Dict[str, Any]]]) -> Union[List[Dict[str, Any]], List[List[Dict[str, Any]]]]:
163
+ """
164
+ Receives request payload, returns predictions.
165
+ """
166
+ if isinstance(data, list):
167
+ return [self._process_single(item) for item in data]
168
+ elif isinstance(data, dict):
169
+ # If payload is {"inputs": ["str1", "str2"], ...}
170
+ if isinstance(data.get("inputs"), list):
171
+ results = []
172
+ for inp in data["inputs"]:
173
+ single_data = dict(data)
174
+ single_data["inputs"] = inp
175
+ results.append(self._process_single(single_data))
176
+ return results
177
+ return self._process_single(data)
178
+ elif isinstance(data, str):
179
+ return self._process_single(data)
180
+ else:
181
+ raise ValueError(f"Unsupported payload type: {type(data)}")
merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:774332cd8b9c738b46ea436523400d01a0a0a3fb54d2cc5296cca178cd8185b8
3
+ size 1204159808
requirements.txt ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ torch>=2.0.0
2
+ transformers>=4.40.0
3
+ safetensors>=0.4.0
4
+ accelerate>=0.28.0
5
+ bitsandbytes>=0.43.0
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
3
+ size 11421892
tokenizer_config.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "clean_up_tokenization_spaces": false,
6
+ "eos_token": "<|im_end|>",
7
+ "errors": "replace",
8
+ "extra_special_tokens": [
9
+ "<|im_start|>",
10
+ "<|im_end|>",
11
+ "<|object_ref_start|>",
12
+ "<|object_ref_end|>",
13
+ "<|box_start|>",
14
+ "<|box_end|>",
15
+ "<|quad_start|>",
16
+ "<|quad_end|>",
17
+ "<|vision_start|>",
18
+ "<|vision_end|>",
19
+ "<|vision_pad|>",
20
+ "<|image_pad|>",
21
+ "<|video_pad|>"
22
+ ],
23
+ "is_local": true,
24
+ "local_files_only": false,
25
+ "model_max_length": 131072,
26
+ "pad_token": "<|endoftext|>",
27
+ "split_special_tokens": false,
28
+ "tokenizer_class": "Qwen2Tokenizer",
29
+ "unk_token": null
30
+ }
vocab.json ADDED
The diff for this file is too large to render. See raw diff