Instructions to use THemidli/applied-ner-stage3-bert-tiny with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use THemidli/applied-ner-stage3-bert-tiny with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("token-classification", model="THemidli/applied-ner-stage3-bert-tiny")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForTokenClassification tokenizer = AutoTokenizer.from_pretrained("THemidli/applied-ner-stage3-bert-tiny") model = AutoModelForTokenClassification.from_pretrained("THemidli/applied-ner-stage3-bert-tiny", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload artifacts/stage3_cpu_benchmark.json with huggingface_hub
Browse files
artifacts/stage3_cpu_benchmark.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"stage": 3,
|
| 3 |
+
"model_path": "models/stage3-bert-tiny-ner",
|
| 4 |
+
"device": "cpu",
|
| 5 |
+
"torch_threads": 8,
|
| 6 |
+
"parameter_count": 4371601,
|
| 7 |
+
"parameter_memory_mb": 17.486404,
|
| 8 |
+
"artifact_size_mb": 18.209407,
|
| 9 |
+
"rss_before_load_mb": 442.44992,
|
| 10 |
+
"rss_after_load_mb": 474.300416,
|
| 11 |
+
"model_load_rss_delta_mb": 31.850496,
|
| 12 |
+
"rss_after_inference_mb": 483.753984,
|
| 13 |
+
"inference_rss_delta_mb": 41.304064,
|
| 14 |
+
"benchmark_examples": 2000,
|
| 15 |
+
"batch_size": 32,
|
| 16 |
+
"elapsed_seconds": 0.1745212919995538,
|
| 17 |
+
"throughput_examples_per_second": 11459.919744377743,
|
| 18 |
+
"mean_latency_ms_per_example_at_batch_size": 0.0872606459997769,
|
| 19 |
+
"scope": "end-to-end fast-tokenizer plus PyTorch CPU forward pass on the 40 wild probes"
|
| 20 |
+
}
|