Text Classification
Transformers
TensorBoard
Safetensors
English
llama
text-generation
llama-factory
full
Generated from Trainer
text-embeddings-inference
Instructions to use Rakancorle1/ThinkGuard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Rakancorle1/ThinkGuard with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="Rakancorle1/ThinkGuard")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Rakancorle1/ThinkGuard") model = AutoModelForCausalLM.from_pretrained("Rakancorle1/ThinkGuard", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -92,7 +92,7 @@ print("\n==================================\n")
|
|
| 92 |
|
| 93 |
## Performance
|
| 94 |
Unlike the other three benchmarks, which solely evaluate Safety Assessment (i.e., binary safe/unsafe classification), BeaverTails is a multi-class classification benchmark. Its F1 score evaluation extends beyond simple Safety Assessment to measure accuracy across multiple risk categories, providing a more fine-grained assessment of model performance.
|
| 95 |
-
|
| 96 |
## Model Description
|
| 97 |
|
| 98 |
- **Model type:** Guardrail model fine-tuned to enhance safety classification with critiques-augmented finetuning.
|
|
|
|
| 92 |
|
| 93 |
## Performance
|
| 94 |
Unlike the other three benchmarks, which solely evaluate Safety Assessment (i.e., binary safe/unsafe classification), BeaverTails is a multi-class classification benchmark. Its F1 score evaluation extends beyond simple Safety Assessment to measure accuracy across multiple risk categories, providing a more fine-grained assessment of model performance.
|
| 95 |
+

|
| 96 |
## Model Description
|
| 97 |
|
| 98 |
- **Model type:** Guardrail model fine-tuned to enhance safety classification with critiques-augmented finetuning.
|