Upload benchmark_config.yaml with huggingface_hub
Browse files- benchmark_config.yaml +56 -0
benchmark_config.yaml
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
model:
|
| 2 |
+
base_url: "http://127.0.0.1:8000"
|
| 3 |
+
max_tokens: 512
|
| 4 |
+
temperature: 0.1
|
| 5 |
+
timeout: 30
|
| 6 |
+
|
| 7 |
+
datasets:
|
| 8 |
+
benchmark_dataset:
|
| 9 |
+
file_path: "ner_benchmark_dataset.jsonl"
|
| 10 |
+
sample_size: 100 # Use first 100 examples for quick benchmarking
|
| 11 |
+
instruction_field: "instruction"
|
| 12 |
+
input_field: "input"
|
| 13 |
+
expected_output_field: "response"
|
| 14 |
+
|
| 15 |
+
metrics:
|
| 16 |
+
# Primary metrics for HuggingFace
|
| 17 |
+
entity_recognition:
|
| 18 |
+
name: "Entity Recognition F1 Score"
|
| 19 |
+
description: "F1 score for named entity recognition accuracy"
|
| 20 |
+
type: "f1"
|
| 21 |
+
|
| 22 |
+
precision:
|
| 23 |
+
name: "Precision Score"
|
| 24 |
+
description: "Precision for entity recognition"
|
| 25 |
+
type: "precision"
|
| 26 |
+
|
| 27 |
+
recall:
|
| 28 |
+
name: "Recall Score"
|
| 29 |
+
description: "Recall for entity recognition"
|
| 30 |
+
type: "recall"
|
| 31 |
+
|
| 32 |
+
latency:
|
| 33 |
+
name: "Average Latency"
|
| 34 |
+
description: "Average response time in milliseconds"
|
| 35 |
+
type: "latency"
|
| 36 |
+
|
| 37 |
+
# Entity type specific performance
|
| 38 |
+
entity_types:
|
| 39 |
+
person:
|
| 40 |
+
name: "Person Entity Recognition"
|
| 41 |
+
keywords: ["PERSON", "person", "Person"]
|
| 42 |
+
organization:
|
| 43 |
+
name: "Organization Entity Recognition"
|
| 44 |
+
keywords: ["ORG", "organization", "Organization"]
|
| 45 |
+
location:
|
| 46 |
+
name: "Location Entity Recognition"
|
| 47 |
+
keywords: ["LOC", "location", "Location"]
|
| 48 |
+
miscellaneous:
|
| 49 |
+
name: "Miscellaneous Entity Recognition"
|
| 50 |
+
keywords: ["MISC", "miscellaneous", "Miscellaneous"]
|
| 51 |
+
|
| 52 |
+
output:
|
| 53 |
+
results_file: "benchmarks.txt"
|
| 54 |
+
detailed_results_file: "benchmark_results.json"
|
| 55 |
+
include_examples: true
|
| 56 |
+
max_examples: 10
|