Hugging Face's logo

BRR1154
/
SLM-ResourceManager-v2

Text Generation
Transformers
Safetensors
English
qwen2
resource-manager
vram-sizing
scheduler
qwen2.5
null-space-lora
slm-v2
conversational
Eval Results (legacy)
text-generation-inference
Model card Files
xet
 Community
BRR1154 commited on
Commit
e6fdeb0
·
verified ·
1 Parent(s): d6b8641

Publish BRR1154/SLM-ResourceManager-v2: verified Swarm-14 v2 Real-Sleep Null-Space checkpoint

Browse files
Files changed (9) hide show
  1. .gitattributes +1 -0
  2. README.md +83 -0
  3. chat_template.jinja +54 -0
  4. config.json +57 -0
  5. generation_config.json +14 -0
  6. model.safetensors +3 -0
  7. real_training_telemetry.json +231 -0
  8. tokenizer.json +3 -0
  9. tokenizer_config.json +30 -0
.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
README.md ADDED
@@ -0,0 +1,83 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: apache-2.0
5
+ library_name: transformers
6
+ pipeline_tag: text-generation
7
+ base_model:
8
+ - Qwen/Qwen2.5-0.5B-Instruct
9
+ - BRR1154/SLM-ResourceManager-v1
10
+ tags:
11
+ - resource-manager
12
+ - vram-sizing
13
+ - scheduler
14
+ - qwen2.5
15
+ - null-space-lora
16
+ - slm-v2
17
+ model-index:
18
+ - name: SLM-ResourceManager-v2
19
+ results:
20
+ - task:
21
+ type: text-generation
22
+ name: VRAM-Sizing-16
23
+ metrics:
24
+ - name: Domain Execution / Solver Accuracy (After v2)
25
+ type: accuracy
26
+ value: 12.5
27
+ - name: Domain Execution / Solver Accuracy (Before v1 Baseline)
28
+ type: accuracy
29
+ value: 0.0
30
+ - task:
31
+ type: multiple-choice
32
+ name: ARC-Easy-25 General Retention Guard
33
+ metrics:
34
+ - name: ARC-Easy-25 Raw / Calibrated (%)
35
+ type: accuracy
36
+ value: 56.0
37
+ ---
38
+
39
+ # SLM-ResourceManager-v2: GPU VRAM & Execution Budget Scheduler (v2)
40
+
41
+ **Lead Architect:** Bhavanam Rajendra Reddy (*Runtime-Slayers* | ORCID: [0009-0005-3349-1162](https://orcid.org/0009-0005-3349-1162))
42
+ **Base v1 Checkpoint:** [`BRR1154/SLM-ResourceManager-v1`](https://huggingface.co/BRR1154/SLM-ResourceManager-v1) (`Qwen/Qwen2.5-0.5B-Instruct`, 494M params, BF16)
43
+ **Training Method:** Real PyTorch `LoRALinear` Backpropagation + Wake-Error `ABSTAIN` Calibration + Empirical Activation-Covariance SVD Null-Space Consolidation ($\Delta W_{\text{proj}} = \Delta W - (\Delta W V_{\text{base}}) V_{\text{base}}^T$)
44
+
45
+ ---
46
+
47
+ ## Verified Before (`v1`) vs After (`v2`) Evaluation Scorecard
48
+
49
+ All metrics below are directly measured on disjoint held-out test suites (`eval_swarm14.py`) with real execution/solver verifiers (`sqlite3`, `z3`, `lxml.html`, `HTMLParser`, `subprocess` Python, Kahn DAG checker). Full weight telemetry is included in `real_training_telemetry.json`.
50
+
51
+ | Metric | `v1` Baseline (`BRR1154/SLM-ResourceManager-v1`) | `v2` Checkpoint (`BRR1154/SLM-ResourceManager-v2`) | Delta |
52
+ | :--- | :---: | :---: | :---: |
53
+ | **Domain (`VRAM-Sizing-16`)** | **0.00%** | **12.50%** | **`+12.50` pts** |
54
+ | **ARC-Easy-25 Retention (Raw / PMI-Cal)** | 64.0% / 72.0% | 56.0% / 56.0% | `-8.0` / `-16.0` pts |
55
+ | **AA-Omniscience-12 Calibration Index** | -25.00 | -8.33 | `+16.67` pts |
56
+
57
+ ---
58
+
59
+ ## Null-Space Weight Surgery Audit
60
+
61
+ - **Targeted Layer Band:** `10-23` (`q_proj, v_proj, o_proj, down_proj`)
62
+ - **Optimization Schedule:** `40` gradient steps (`LoRA rank=16`, empirical SVD subspace `k=24`)
63
+ - **Mutated Projections Verified:** `56/56` modules (`sha256` verified before vs after merge)
64
+ - **Max Frobenius Shift Ratio:** `1.532%`
65
+ - **Max Null-Space Orthogonality Error:** `2.648e-09`
66
+
67
+ ---
68
+
69
+ ## Quickstart (`transformers`)
70
+
71
+ ```python
72
+ import torch
73
+ from transformers import AutoModelForCausalLM, AutoTokenizer
74
+
75
+ model_id = "BRR1154/SLM-ResourceManager-v2"
76
+ tokenizer = AutoTokenizer.from_pretrained(model_id)
77
+ model = AutoModelForCausalLM.from_pretrained(model_id, dtype=torch.bfloat16, device_map="auto")
78
+
79
+ messages = [{"role": "user", "content": "Solve the specialist task concisely and deterministically."}]
80
+ inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to(model.device)
81
+ out = model.generate(inputs, max_new_tokens=160, do_sample=False)
82
+ print(tokenizer.decode(out[0][inputs.shape[1]:], skip_special_tokens=True))
83
+ ```
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 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 896,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 4864,
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
+ ],
39
+ "max_position_embeddings": 32768,
40
+ "max_window_layers": 21,
41
+ "model_type": "qwen2",
42
+ "num_attention_heads": 14,
43
+ "num_hidden_layers": 24,
44
+ "num_key_value_heads": 2,
45
+ "pad_token_id": null,
46
+ "rms_norm_eps": 1e-06,
47
+ "rope_parameters": {
48
+ "rope_theta": 1000000.0,
49
+ "rope_type": "default"
50
+ },
51
+ "sliding_window": null,
52
+ "tie_word_embeddings": true,
53
+ "transformers_version": "5.16.1",
54
+ "use_cache": true,
55
+ "use_sliding_window": false,
56
+ "vocab_size": 151936
57
+ }
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": "5.16.1"
14
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:200aaa787837e8178dd9e8d703c3e3fbc5d5a93d169f7cd5f76ff54f452256f0
3
+ size 988097824
real_training_telemetry.json ADDED
@@ -0,0 +1,231 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_tag": "SLM-ResourceManager-v1-v2",
3
+ "base_path": "C:\\Users\\brr33\\.gemini\\antigravity\\scratch\\intriq\\slm_swarm_engine\\models\\SLM-ResourceManager-v1",
4
+ "output_dir": "C:\\Users\\brr33\\.gemini\\antigravity\\scratch\\intriq\\real_sleep_trainer\\checkpoints_v2\\SLM-ResourceManager-v1-v2",
5
+ "curriculum_path": "C:\\Users\\brr33\\.gemini\\antigravity\\scratch\\intriq\\real_sleep_trainer\\swarm14\\curricula\\SLM-ResourceManager-v1.json",
6
+ "layer_band": "10-23",
7
+ "save_bf16": true,
8
+ "steps": 40,
9
+ "grad_accum": 2,
10
+ "effective_examples": 160,
11
+ "lora_rank": 16,
12
+ "nullspace_k": 24,
13
+ "frobenius_budget": 0.14,
14
+ "elapsed_total_sec": 488.88,
15
+ "post_null_option_logprobs": {
16
+ "A": -15.421,
17
+ "B": -15.0,
18
+ "C": -14.536,
19
+ "D": -14.522,
20
+ "E": -16.089
21
+ },
22
+ "loss_history": [
23
+ {
24
+ "step": 1,
25
+ "sft_loss": 2.2221,
26
+ "mcq_loss": 0.0607,
27
+ "prior_kl": 0.85201
28
+ },
29
+ {
30
+ "step": 10,
31
+ "sft_loss": 0.0208,
32
+ "mcq_loss": 1.4402,
33
+ "prior_kl": 0.09041
34
+ },
35
+ {
36
+ "step": 20,
37
+ "sft_loss": 0.2937,
38
+ "mcq_loss": 0.8394,
39
+ "prior_kl": 0.03171
40
+ },
41
+ {
42
+ "step": 30,
43
+ "sft_loss": 0.1364,
44
+ "mcq_loss": 0.4799,
45
+ "prior_kl": 0.01281
46
+ },
47
+ {
48
+ "step": 40,
49
+ "sft_loss": 0.9718,
50
+ "mcq_loss": 0.1196,
51
+ "prior_kl": 0.02133
52
+ }
53
+ ],
54
+ "nullspace_merge_audit": {
55
+ "modules_merged": 56,
56
+ "modules_hash_changed": 56,
57
+ "nullspace_projected": true,
58
+ "max_orthogonality_error": 2.648448571562767e-09,
59
+ "max_frobenius_ratio_pct": 1.532,
60
+ "layer_sample": [
61
+ {
62
+ "module": "model.layers.10.self_attn.q_proj",
63
+ "shape": [
64
+ 896,
65
+ 896
66
+ ],
67
+ "w0_frob": 19.2065,
68
+ "dw_frob": 0.23478,
69
+ "frob_ratio_pct": 1.222,
70
+ "nullspace_ortho_err": 1.7898855730891228e-09,
71
+ "clamped": false,
72
+ "sha256_before": "46f7bf7e1965bde1",
73
+ "sha256_after": "2e1e947256a307f8"
74
+ },
75
+ {
76
+ "module": "model.layers.10.self_attn.v_proj",
77
+ "shape": [
78
+ 128,
79
+ 896
80
+ ],
81
+ "w0_frob": 6.8381,
82
+ "dw_frob": 0.0811,
83
+ "frob_ratio_pct": 1.186,
84
+ "nullspace_ortho_err": 1.6880221664905548e-09,
85
+ "clamped": false,
86
+ "sha256_before": "a4bedd8404473749",
87
+ "sha256_after": "fe8ed096eb20f347"
88
+ },
89
+ {
90
+ "module": "model.layers.10.self_attn.o_proj",
91
+ "shape": [
92
+ 896,
93
+ 896
94
+ ],
95
+ "w0_frob": 16.2924,
96
+ "dw_frob": 0.22735,
97
+ "frob_ratio_pct": 1.395,
98
+ "nullspace_ortho_err": 9.058567229658365e-10,
99
+ "clamped": false,
100
+ "sha256_before": "1870c804cbb395b2",
101
+ "sha256_after": "4b2d43b76fa48fe8"
102
+ },
103
+ {
104
+ "module": "model.layers.10.mlp.down_proj",
105
+ "shape": [
106
+ 896,
107
+ 4864
108
+ ],
109
+ "w0_frob": 36.6858,
110
+ "dw_frob": 0.22735,
111
+ "frob_ratio_pct": 0.62,
112
+ "nullspace_ortho_err": 5.384208634495735e-10,
113
+ "clamped": false,
114
+ "sha256_before": "e8fccd08b4da91fb",
115
+ "sha256_after": "8516bbfbfdf2c73f"
116
+ },
117
+ {
118
+ "module": "model.layers.11.self_attn.q_proj",
119
+ "shape": [
120
+ 896,
121
+ 896
122
+ ],
123
+ "w0_frob": 18.4899,
124
+ "dw_frob": 0.24435,
125
+ "frob_ratio_pct": 1.322,
126
+ "nullspace_ortho_err": 1.7025740817189217e-09,
127
+ "clamped": false,
128
+ "sha256_before": "000ae2d2ac3f6bab",
129
+ "sha256_after": "f076cdea4b712de6"
130
+ },
131
+ {
132
+ "module": "model.layers.11.self_attn.v_proj",
133
+ "shape": [
134
+ 128,
135
+ 896
136
+ ],
137
+ "w0_frob": 5.7194,
138
+ "dw_frob": 0.08606,
139
+ "frob_ratio_pct": 1.505,
140
+ "nullspace_ortho_err": 9.604264050722122e-10,
141
+ "clamped": false,
142
+ "sha256_before": "e8b86d80f07b6700",
143
+ "sha256_after": "fd1ddfcb2e9c22c4"
144
+ },
145
+ {
146
+ "module": "model.layers.11.self_attn.o_proj",
147
+ "shape": [
148
+ 896,
149
+ 896
150
+ ],
151
+ "w0_frob": 15.6306,
152
+ "dw_frob": 0.22839,
153
+ "frob_ratio_pct": 1.461,
154
+ "nullspace_ortho_err": 1.5934347175061703e-09,
155
+ "clamped": false,
156
+ "sha256_before": "7cbd60e9e1311cb8",
157
+ "sha256_after": "b5341056e9dc25c9"
158
+ },
159
+ {
160
+ "module": "model.layers.11.mlp.down_proj",
161
+ "shape": [
162
+ 896,
163
+ 4864
164
+ ],
165
+ "w0_frob": 37.3923,
166
+ "dw_frob": 0.22568,
167
+ "frob_ratio_pct": 0.604,
168
+ "nullspace_ortho_err": 2.9740476747974753e-10,
169
+ "clamped": false,
170
+ "sha256_before": "40a802e80b994e03",
171
+ "sha256_after": "2a62885d3a20e9f9"
172
+ },
173
+ {
174
+ "module": "model.layers.12.self_attn.q_proj",
175
+ "shape": [
176
+ 896,
177
+ 896
178
+ ],
179
+ "w0_frob": 19.277,
180
+ "dw_frob": 0.24853,
181
+ "frob_ratio_pct": 1.289,
182
+ "nullspace_ortho_err": 1.2150849215686321e-09,
183
+ "clamped": false,
184
+ "sha256_before": "d1657c07c6fddd0b",
185
+ "sha256_after": "9b03a7a4cb8b3eaa"
186
+ },
187
+ {
188
+ "module": "model.layers.12.self_attn.v_proj",
189
+ "shape": [
190
+ 128,
191
+ 896
192
+ ],
193
+ "w0_frob": 6.3311,
194
+ "dw_frob": 0.0884,
195
+ "frob_ratio_pct": 1.396,
196
+ "nullspace_ortho_err": 1.4915713109076023e-09,
197
+ "clamped": false,
198
+ "sha256_before": "fdb34b5d85d83079",
199
+ "sha256_after": "360afab1d6eebd06"
200
+ },
201
+ {
202
+ "module": "model.layers.12.self_attn.o_proj",
203
+ "shape": [
204
+ 896,
205
+ 896
206
+ ],
207
+ "w0_frob": 16.0116,
208
+ "dw_frob": 0.22488,
209
+ "frob_ratio_pct": 1.404,
210
+ "nullspace_ortho_err": 1.2951204553246498e-09,
211
+ "clamped": false,
212
+ "sha256_before": "a54a85a46eae3759",
213
+ "sha256_after": "59062925e69db48c"
214
+ },
215
+ {
216
+ "module": "model.layers.12.mlp.down_proj",
217
+ "shape": [
218
+ 896,
219
+ 4864
220
+ ],
221
+ "w0_frob": 38.0378,
222
+ "dw_frob": 0.22034,
223
+ "frob_ratio_pct": 0.579,
224
+ "nullspace_ortho_err": 3.3378455555066466e-10,
225
+ "clamped": false,
226
+ "sha256_before": "6bcde15bf26b0c41",
227
+ "sha256_after": "fff053ad32e002c3"
228
+ }
229
+ ]
230
+ }
231
+ }
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
+ }