Update README.md
Browse files
README.md
CHANGED
@@ -173,7 +173,7 @@ chat()
|
|
173 |
- **Model type:** Text-generation
|
174 |
- **Language(s) (NLP):** English
|
175 |
- **License:** apache-2.0
|
176 |
-
- **Finetuned from model
|
177 |
|
178 |
## Bias, Risks, and Limitations
|
179 |
|
@@ -183,6 +183,18 @@ chat()
|
|
183 |
* Over-Reliance Risk - Outputs could inadvertently worsen symptoms if users interpret them as definitive advice.
|
184 |
* Intended Use - Assist with general emotional support, not diagnosis or treatment.
|
185 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
186 |
### Model Sources [optional]
|
187 |
|
188 |
<!-- Provide the basic links for the model. -->
|
|
|
173 |
- **Model type:** Text-generation
|
174 |
- **Language(s) (NLP):** English
|
175 |
- **License:** apache-2.0
|
176 |
+
- **Finetuned from model:** falcon-rw-1b
|
177 |
|
178 |
## Bias, Risks, and Limitations
|
179 |
|
|
|
183 |
* Over-Reliance Risk - Outputs could inadvertently worsen symptoms if users interpret them as definitive advice.
|
184 |
* Intended Use - Assist with general emotional support, not diagnosis or treatment.
|
185 |
|
186 |
+
## Training Hyperparameters
|
187 |
+
|
188 |
+
| Hyperparameter | Value |
|
189 |
+
| ------------- | ------------- |
|
190 |
+
| Precision | float16 |
|
191 |
+
| Optimizer | AdamW_32bit |
|
192 |
+
| Learning rate | 2e-4 |
|
193 |
+
| Weight decay | 1e-2 |
|
194 |
+
| Batch size | 1 |
|
195 |
+
| Training Epochs | 3 |
|
196 |
+
| Quantization | 8-Bit |
|
197 |
+
|
198 |
### Model Sources [optional]
|
199 |
|
200 |
<!-- Provide the basic links for the model. -->
|