Upload 3 files

README.md

@@ -1,12 +1,2 @@
----
-title: Mathstral Assistant
-emoji: 😻
-colorFrom: purple
-colorTo: gray
-sdk: streamlit
-sdk_version: 1.36.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# mathstral_assistant
+# Mathstral-7B Assistant for Physics, Engineering, and Mathematics  This Streamlit app provides an interface for the Mathstral-7B model, assisting with questions in physics, engineering, and mathematics.  ## Setup and Running  1. Clone this repository:

app.py

@@ -0,0 +1,22 @@
+import streamlit as st
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+# Load the model and tokenizer
+model_name = "mistralai/mathstral-7B-v0.1"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+
+st.title("Mathstral-7B Assistant for Physics, Engineering, and Mathematics")
+
+st.write("Enter your question related to physics, engineering, or mathematics:")
+prompt = st.text_area("Question:", "Example: Calculate the force on an object with mass 10 kg and acceleration 2 m/s².")
+
+max_length = st.sidebar.slider("Max Response Length", min_value=10, max_value=200, value=100)
+
+if st.button("Generate Response"):
+    with st.spinner("Generating response..."):
+        full_prompt = f"Assistant in physics, engineering, and mathematics: {prompt}"
+        inputs = tokenizer(full_prompt, return_tensors="pt")
+        outputs = model.generate(**inputs, max_length=max_length)
+        response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+        st.write("Response:", response)

requirements.txt

@@ -0,0 +1,3 @@
+streamlit
+transformers
+torch