baban commited on
Commit
dcf5088
·
verified ·
1 Parent(s): 8dac678

Upload perturb_042_seed_1807519518_sigma_0p0005/validation/cache/reports/granite-4.0-8b-preview-r260310a/bfcl_v4.json with huggingface_hub

Browse files
perturb_042_seed_1807519518_sigma_0p0005/validation/cache/reports/granite-4.0-8b-preview-r260310a/bfcl_v4.json ADDED
@@ -0,0 +1,188 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "granite-4.0-8b-preview-r260310a@bfcl_v4",
3
+ "dataset_name": "bfcl_v4",
4
+ "dataset_pretty_name": "BFCL-v4",
5
+ "dataset_description": "\n## Overview\n\nBFCL-v4 (Berkeley Function-Calling Leaderboard V4) is a comprehensive benchmark for evaluating agentic function-calling capabilities of LLMs. It tests web search, memory operations, and format sensitivity as building blocks for agentic applications.\n\n## Task Description\n\n- **Task Type**: Agentic Function Calling Evaluation\n- **Input**: Scenarios requiring function calls for web search, memory, etc.\n- **Output**: Correct function calls with proper arguments\n- **Capabilities**: Web search, memory read/write, format handling\n\n## Key Features\n\n- Holistic agentic evaluation framework\n- Tests building blocks for LLM agents (search, memory, formatting)\n- Multiple scoring categories for detailed analysis\n- Supports both FC models and non-FC models\n- Optional web search with SerpAPI integration\n\n## Evaluation Notes\n\n- **Installation Required**: `pip install bfcl-eval==2025.10.27.1`\n- Primary metric: **Accuracy** per category\n- Configure `is_fc_model` based on model capabilities\n- Optional: Set `SERPAPI_API_KEY` for web search tasks\n- [Usage Example](https://evalscope.readthedocs.io/en/latest/third_party/bfcl_v4.html)\n",
6
+ "model_name": "granite-4.0-8b-preview-r260310a",
7
+ "score": 0.5269,
8
+ "metrics": [
9
+ {
10
+ "name": "acc",
11
+ "num": 2190,
12
+ "score": 0.5269,
13
+ "macro_score": 0.5269,
14
+ "categories": [
15
+ {
16
+ "name": [
17
+ "default"
18
+ ],
19
+ "num": 2190,
20
+ "score": 0.5269,
21
+ "macro_score": 0.4705,
22
+ "subsets": [
23
+ {
24
+ "name": "irrelevance",
25
+ "score": 0.8208,
26
+ "num": 240,
27
+ "is_aggregate": false
28
+ },
29
+ {
30
+ "name": "multi_turn_base",
31
+ "score": 0.08,
32
+ "num": 200,
33
+ "is_aggregate": false
34
+ },
35
+ {
36
+ "name": "multi_turn_long_context",
37
+ "score": 0.045,
38
+ "num": 200,
39
+ "is_aggregate": false
40
+ },
41
+ {
42
+ "name": "multi_turn_miss_func",
43
+ "score": 0.005,
44
+ "num": 200,
45
+ "is_aggregate": false
46
+ },
47
+ {
48
+ "name": "multi_turn_miss_param",
49
+ "score": 0.04,
50
+ "num": 200,
51
+ "is_aggregate": false
52
+ },
53
+ {
54
+ "name": "multiple",
55
+ "score": 0.915,
56
+ "num": 200,
57
+ "is_aggregate": false
58
+ },
59
+ {
60
+ "name": "parallel",
61
+ "score": 0.775,
62
+ "num": 200,
63
+ "is_aggregate": false
64
+ },
65
+ {
66
+ "name": "parallel_multiple",
67
+ "score": 0.81,
68
+ "num": 200,
69
+ "is_aggregate": false
70
+ },
71
+ {
72
+ "name": "simple_java",
73
+ "score": 0.37,
74
+ "num": 100,
75
+ "is_aggregate": false
76
+ },
77
+ {
78
+ "name": "simple_javascript",
79
+ "score": 0.4,
80
+ "num": 50,
81
+ "is_aggregate": false
82
+ },
83
+ {
84
+ "name": "simple_python",
85
+ "score": 0.915,
86
+ "num": 400,
87
+ "is_aggregate": false
88
+ }
89
+ ]
90
+ },
91
+ {
92
+ "name": [
93
+ "-"
94
+ ],
95
+ "num": 0,
96
+ "score": 0.0,
97
+ "macro_score": 0.0,
98
+ "subsets": [
99
+ {
100
+ "name": "MULTI_TURN",
101
+ "score": 0.0425,
102
+ "num": 800,
103
+ "is_aggregate": true
104
+ },
105
+ {
106
+ "name": "NON_LIVE",
107
+ "score": 0.7654,
108
+ "num": 1150,
109
+ "is_aggregate": true
110
+ },
111
+ {
112
+ "name": "HALLUCINATION",
113
+ "score": 0.8208,
114
+ "num": 240,
115
+ "is_aggregate": true
116
+ },
117
+ {
118
+ "name": "OVERALL",
119
+ "score": 0.1714,
120
+ "num": 2190,
121
+ "is_aggregate": true
122
+ }
123
+ ]
124
+ }
125
+ ]
126
+ }
127
+ ],
128
+ "analysis": "N/A",
129
+ "perf_metrics": {
130
+ "summary": {
131
+ "n_samples": 7982,
132
+ "latency": {
133
+ "mean": 2.380037,
134
+ "std": 19.375791,
135
+ "min": 0.063361,
136
+ "25%": 0.720828,
137
+ "50%": 1.189623,
138
+ "75%": 2.040548,
139
+ "90%": 3.504383,
140
+ "99%": 9.735258,
141
+ "max": 760.008047
142
+ },
143
+ "throughput": {
144
+ "avg_output_tps": 38.85,
145
+ "avg_req_ps": 0.4202
146
+ },
147
+ "usage": {
148
+ "input_tokens": {
149
+ "mean": 6603.556502,
150
+ "std": 5309.918046,
151
+ "min": 237.0,
152
+ "25%": 4590.5,
153
+ "50%": 6944.0,
154
+ "75%": 8598.0,
155
+ "90%": 9963.0,
156
+ "99%": 38482.28,
157
+ "max": 87327.0
158
+ },
159
+ "output_tokens": {
160
+ "mean": 92.463668,
161
+ "std": 249.789082,
162
+ "min": 1.0,
163
+ "25%": 35.0,
164
+ "50%": 57.0,
165
+ "75%": 104.0,
166
+ "90%": 169.0,
167
+ "99%": 486.19,
168
+ "max": 10785.0
169
+ },
170
+ "total_tokens": {
171
+ "mean": 6696.02017,
172
+ "std": 5330.221586,
173
+ "min": 247.0,
174
+ "25%": 4683.5,
175
+ "50%": 7016.0,
176
+ "75%": 8682.75,
177
+ "90%": 10049.8,
178
+ "99%": 38638.07,
179
+ "max": 87562.0
180
+ },
181
+ "total_input_tokens": 52709588,
182
+ "total_output_tokens": 738045,
183
+ "total_tokens_count": 53447633
184
+ }
185
+ }
186
+ },
187
+ "num": 2190
188
+ }