Update README.md

README.md

@@ -1,199 +1,169 @@
 ---
 library_name: transformers
-tags: []
+tags: [Text Generation, Question-Answering]
 ---
 
 # Model Card for Model ID
 
 <!-- Provide a quick summary of what the model is/does. -->
 
+It's a fine tuned version of Phi-2 model by Microsoft on Amod/mental_health_counseling_conversations.
 
 
-## Model Details
-
-### Model Description
-
-<!-- Provide a longer summary of what this model is. -->
-
-This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
-
-- **Developed by:** [More Information Needed]
-- **Funded by [optional]:** [More Information Needed]
-- **Shared by [optional]:** [More Information Needed]
-- **Model type:** [More Information Needed]
-- **Language(s) (NLP):** [More Information Needed]
-- **License:** [More Information Needed]
-- **Finetuned from model [optional]:** [More Information Needed]
-
-### Model Sources [optional]
-
-<!-- Provide the basic links for the model. -->
-
-- **Repository:** [More Information Needed]
-- **Paper [optional]:** [More Information Needed]
-- **Demo [optional]:** [More Information Needed]
-
 ## Uses
 
 <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+The above model, with applicable changes to the generation_config file passed to model.generate() fucntion can lead to generation of better resulst which could then be used for Health Counseling Chatbot dev.
 
-### Direct Use
-
-<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
-
-[More Information Needed]
-
-### Downstream Use [optional]
-
-<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
-
-[More Information Needed]
-
-### Out-of-Scope Use
 
-<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
-
-[More Information Needed]
 
 ## Bias, Risks, and Limitations
 
-<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+The model was developed as a proof-of-concept type hobby project and is not intended to be used without careful considerations of its implications.
 
 [More Information Needed]
 
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
 
 ## How to Get Started with the Model
 
 Use the code below to get started with the model.
 
-[More Information Needed]
+```python
 
-## Training Details
+new_model = "YuvrajSingh9886/medicinal-QnA-phi2-custom"
 
-### Training Data
+prompt = "I have been feeling more and more down for over a month. I have started having trouble sleeping due to panic attacks, but they are almost never triggered by something that I know of."
 
-<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+tokens = tokenizer(f"### Question: {prompt}", return_tensors='pt').to('cuda')
+tokenizer.pad_token = tokenizer.eos_token
+outputs = model.generate(**tokens, max_new_tokens=1024, num_beams=5,
+    no_repeat_ngram_size=2,
+    early_stopping=True
+                         )
+print(tokenizer.batch_decode(outputs,skip_special_tokens=True)[0])
 
-[More Information Needed]
+```
 
-### Training Procedure
-
-<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
-
-#### Preprocessing [optional]
-
-[More Information Needed]
+## Training Details
 
+### Training Data
 
-#### Training Hyperparameters
 
-- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+#### Hardware
 
-#### Speeds, Sizes, Times [optional]
+Epcohs: 10
+Hardware: (1) RTX 4090 (24GB VRAM) 48GB 8vCPU (RAM)
+          Hard Disk: 40GB
 
-<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
 
 [More Information Needed]
 
-## Evaluation
-
-<!-- This section describes the evaluation protocols and provides the results. -->
-
-### Testing Data, Factors & Metrics
-
-#### Testing Data
-
-<!-- This should link to a Dataset Card if possible. -->
-
-[More Information Needed]
+### Training Procedure
 
-#### Factors
+QLoRA was used for quantization purposes.
 
-<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+Phi-2 model from Huggingface with BitsandBytes support
 
-[More Information Needed]
 
-#### Metrics
+#### Preprocessing [optional]
 
-<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+```python
 
-[More Information Needed]
+def format_phi2(row):
+    question = row['Context']
+    answer = row['Response']
 
-### Results
+#     text = f"[INST] {question} [/INST] {answer}".replace('\xa0', ' ')
+    text = f"### Question: {question}\n ### Answer: {answer}"
 
-[More Information Needed]
+    return text
+```
 
-#### Summary
+#### Training Hyperparameters
 
 
+LoRA config-
+```bash
+# LoRA attention dimension (int)
+lora_r = 64
 
-## Model Examination [optional]
+# Alpha parameter for LoRA scaling (int)
+lora_alpha = 16
 
-<!-- Relevant interpretability work for the model goes here -->
+# Dropout probability for LoRA layers (float)
+lora_dropout = 0.05
 
-[More Information Needed]
+# Bias (string)
+bias = "none"
 
-## Environmental Impact
+# Task type (string)
+task_type = "CAUSAL_LM"
 
-<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+# Random seed (int)
+seed = 33
+```
 
-Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+Phi-2 config-
 
-- **Hardware Type:** [More Information Needed]
-- **Hours used:** [More Information Needed]
-- **Cloud Provider:** [More Information Needed]
-- **Compute Region:** [More Information Needed]
-- **Carbon Emitted:** [More Information Needed]
+```bash
+# Batch size per GPU for training (int)
+per_device_train_batch_size = 6
 
-## Technical Specifications [optional]
+# Number of update steps to accumulate the gradients for (int)
+gradient_accumulation_steps = 2
 
-### Model Architecture and Objective
+# Initial learning rate (AdamW optimizer) (float)
+learning_rate = 2e-4
 
-[More Information Needed]
+# Optimizer to use (string)
+optim = "paged_adamw_8bit"
 
-### Compute Infrastructure
+# Number of training epochs (int)
+num_train_epochs = 4
 
-[More Information Needed]
+# Linear warmup steps from 0 to learning_rate (int)
+warmup_steps = 10
 
-#### Hardware
+# Enable fp16/bf16 training (set bf16 to True with an A100) (bool)
+fp16 = True
 
-[More Information Needed]
+# Log every X updates steps (int)
+logging_steps = 100
 
-#### Software
+#L2 regularization(prevents overfitting)
+weight_decay=0.0
 
-[More Information Needed]
+#Checkpoint saves
+save_strategy="epoch"
+```
 
-## Citation [optional]
+BnB config
 
-<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+```bash
+# Activate 4-bit precision base model loading (bool)
+load_in_4bit = True
 
-**BibTeX:**
+# Activate nested quantization for 4-bit base models (double quantization) (bool)
+bnb_4bit_use_double_quant = True
 
-[More Information Needed]
+# Quantization type (fp4 or nf4) (string)
+bnb_4bit_quant_type = "nf4"
 
-**APA:**
+# Compute data type for 4-bit base models
+bnb_4bit_compute_dtype = torch.bfloat16
 
-[More Information Needed]
+```
 
-## Glossary [optional]
+### Results
 
-<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+Training loss: 2.229
+Validation loss: 2.223
 
-[More Information Needed]
 
 ## More Information [optional]
 
-[More Information Needed]
+[Phi-2](https://huggingface.co/microsoft/phi-2)
 
 ## Model Card Authors [optional]
 
-[More Information Needed]
-
-## Model Card Contact
-
-[More Information Needed]
+[YuvrajSingh9886](https://huggingface.co/YuvrajSingh9886)