Create app.py

app.py

@@ -0,0 +1,99 @@
+import numpy as np
+import gradio as gr
+
+# ======== LOAD DATASET ========
+with open("dataset.txt", "r", encoding="utf-8") as f:
+    text = f.read().lower()
+
+chars = sorted(list(set(text)))
+vocab_size = len(chars)
+stoi = {ch:i for i,ch in enumerate(chars)}
+itos = {i:ch for i,ch in enumerate(chars)}
+
+def encode(s): return [stoi[c] for c in s]
+def decode(l): return "".join([itos[i] for i in l])
+
+data = np.array(encode(text), dtype=np.int32)
+
+# ======== MODEL SETUP ========
+n_hidden = 64
+Wxh = np.random.randn(n_hidden, vocab_size) * 0.01
+Whh = np.random.randn(n_hidden, n_hidden) * 0.01
+Why = np.random.randn(vocab_size, n_hidden) * 0.01
+bh = np.zeros((n_hidden, 1))
+by = np.zeros((vocab_size, 1))
+
+def softmax(x):
+    e = np.exp(x - np.max(x))
+    return e / np.sum(e)
+
+# ======== TRAINING (1x buat demo, simple) ========
+lr = 1e-1
+seq_len = 25
+for epoch in range(1):  # ⚠️ Cuma 1 epoch biar cepat
+    idx = np.random.randint(0, len(data)-seq_len-1)
+    inputs = data[idx:idx+seq_len]
+    targets = data[idx+1:idx+seq_len+1]
+
+    hs = {-1: np.zeros((n_hidden,1))}
+    loss = 0
+    xs, ys, ps = {}, {}, {}
+    for t in range(seq_len):
+        xs[t] = np.zeros((vocab_size,1))
+        xs[t][inputs[t]] = 1
+        hs[t] = np.tanh(np.dot(Wxh, xs[t]) + np.dot(Whh, hs[t-1]) + bh)
+        ys[t] = np.dot(Why, hs[t]) + by
+        ps[t] = softmax(ys[t])
+        loss += -np.log(ps[t][targets[t],0])
+
+    # Backward
+    dWxh, dWhh, dWhy = np.zeros_like(Wxh), np.zeros_like(Whh), np.zeros_like(Why)
+    dbh, dby = np.zeros_like(bh), np.zeros_like(by)
+    dhnext = np.zeros_like(hs[0])
+    for t in reversed(range(seq_len)):
+        dy = np.copy(ps[t])
+        dy[targets[t]] -= 1
+        dWhy += np.dot(dy, hs[t].T)
+        dby += dy
+        dh = np.dot(Why.T, dy) + dhnext
+        dhraw = (1 - hs[t] * hs[t]) * dh
+        dbh += dhraw
+        dWxh += np.dot(dhraw, xs[t].T)
+        dWhh += np.dot(dhraw, hs[t-1].T)
+        dhnext = np.dot(Whh.T, dhraw)
+
+    for param, dparam in zip([Wxh, Whh, Why, bh, by],
+                              [dWxh, dWhh, dWhy, dbh, dby]):
+        param -= lr * dparam
+
+# ======== GENERATE FUNCTION ========
+def generate(seed, length):
+    h = np.zeros((n_hidden,1))
+    x = np.zeros((vocab_size,1))
+    for c in seed:
+        if c not in stoi:
+            continue
+        x = np.zeros((vocab_size,1))
+        x[stoi[c]] = 1
+        h = np.tanh(np.dot(Wxh, x) + np.dot(Whh, h) + bh)
+
+    out = seed
+    for _ in range(length):
+        y = np.dot(Why, h) + by
+        p = softmax(y)
+        ix = np.random.choice(range(vocab_size), p=p.ravel())
+        x = np.zeros((vocab_size,1))
+        x[ix] = 1
+        h = np.tanh(np.dot(Wxh, x) + np.dot(Whh, h) + bh)
+        out += itos[ix]
+    return out
+
+# ======== GRADIO INTERFACE ========
+gr.Interface(fn=generate,
+             inputs=[
+                 gr.Textbox(label="Seed Text", value="halo "),
+                 gr.Slider(20, 200, value=100, label="Generate Length")
+             ],
+             outputs="text",
+             title="Character-Level Text Generator (RNN)"
+).launch()