Spaces:
Sleeping
Sleeping
Update main.py
Browse files
main.py
CHANGED
|
@@ -4,24 +4,32 @@ from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
|
|
| 4 |
from peft import PeftModel
|
| 5 |
|
| 6 |
# -----------------------------
|
| 7 |
-
# PAGE CONFIG
|
| 8 |
# -----------------------------
|
| 9 |
st.set_page_config(page_title="Kurdish Translator", layout="wide")
|
| 10 |
|
| 11 |
-
|
| 12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
bg = "#111111" if THEME == "dark" else "#ffffff"
|
| 14 |
fg = "#ffffff" if THEME == "dark" else "#000000"
|
| 15 |
card_bg = "rgba(255,255,255,0.07)" if THEME == "dark" else "rgba(0,0,0,0.05)"
|
| 16 |
|
| 17 |
st.markdown(
|
| 18 |
-
f"""
|
| 19 |
-
<style>
|
| 20 |
body {{
|
| 21 |
background-color: {bg};
|
| 22 |
color: {fg};
|
| 23 |
}}
|
| 24 |
-
|
| 25 |
.dots-loader {{
|
| 26 |
display: flex;
|
| 27 |
justify-content: center;
|
|
@@ -37,31 +45,28 @@ body {{
|
|
| 37 |
}}
|
| 38 |
.dots-loader div:nth-child(2) {{ animation-delay: 0.2s; }}
|
| 39 |
.dots-loader div:nth-child(3) {{ animation-delay: 0.4s; }}
|
| 40 |
-
|
| 41 |
@keyframes bounce {{
|
| 42 |
from {{ transform: translateY(0); }}
|
| 43 |
to {{ transform: translateY(-12px); }}
|
| 44 |
}}
|
| 45 |
-
|
| 46 |
.output-box {{
|
| 47 |
padding: 20px;
|
| 48 |
border-radius: 12px;
|
| 49 |
background: {card_bg};
|
| 50 |
font-size: 18px;
|
| 51 |
min-height: 120px;
|
|
|
|
| 52 |
}}
|
| 53 |
-
</style>
|
| 54 |
-
""",
|
| 55 |
unsafe_allow_html=True
|
| 56 |
)
|
| 57 |
|
| 58 |
# -----------------------------
|
| 59 |
-
# LOAD MODEL
|
| 60 |
# -----------------------------
|
| 61 |
BASE_MODEL = "facebook/nllb-200-distilled-600M"
|
| 62 |
LORA_REPO = "junaid17/nllb-kurdish-lora"
|
| 63 |
|
| 64 |
-
|
| 65 |
@st.cache_resource
|
| 66 |
def load_model():
|
| 67 |
tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
|
|
@@ -69,10 +74,8 @@ def load_model():
|
|
| 69 |
model = PeftModel.from_pretrained(base, LORA_REPO)
|
| 70 |
return model, tokenizer
|
| 71 |
|
| 72 |
-
|
| 73 |
model, tokenizer = load_model()
|
| 74 |
|
| 75 |
-
|
| 76 |
# -----------------------------
|
| 77 |
# TRANSLATION FUNCTION
|
| 78 |
# -----------------------------
|
|
@@ -87,6 +90,21 @@ def translate(text, src, tgt):
|
|
| 87 |
)
|
| 88 |
return translator(text)[0]["translation_text"]
|
| 89 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 90 |
|
| 91 |
# -----------------------------
|
| 92 |
# UI
|
|
@@ -96,55 +114,67 @@ st.title("Kurdish β English Translator (NLLB + LoRA)")
|
|
| 96 |
col1, colSwap, colTheme, col2 = st.columns([1, 0.6, 0.8, 1])
|
| 97 |
|
| 98 |
with col1:
|
| 99 |
-
src_lang = st.selectbox(
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
|
| 101 |
with colSwap:
|
| 102 |
-
|
| 103 |
|
| 104 |
with colTheme:
|
| 105 |
-
|
|
|
|
| 106 |
|
| 107 |
with col2:
|
| 108 |
-
tgt_lang = st.selectbox(
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
st.rerun()
|
| 114 |
-
|
| 115 |
-
# Toggle theme
|
| 116 |
-
if toggle:
|
| 117 |
-
st.session_state["theme"] = "dark" if THEME == "light" else "light"
|
| 118 |
-
st.rerun()
|
| 119 |
|
| 120 |
lang_codes = {
|
| 121 |
"English": "eng_Latn",
|
| 122 |
"Kurdish": "ckb_Arab"
|
| 123 |
}
|
| 124 |
|
| 125 |
-
src_code = lang_codes[src_lang]
|
| 126 |
-
tgt_code = lang_codes[tgt_lang]
|
| 127 |
|
| 128 |
text = st.text_area("Enter text:", height=180)
|
| 129 |
|
| 130 |
-
|
| 131 |
# -----------------------------
|
| 132 |
-
# TRANSLATE BUTTON
|
| 133 |
# -----------------------------
|
| 134 |
-
if st.button("Translate"):
|
| 135 |
if not text.strip():
|
| 136 |
st.warning("Please enter some text.")
|
| 137 |
else:
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
from peft import PeftModel
|
| 5 |
|
| 6 |
# -----------------------------
|
| 7 |
+
# PAGE CONFIG
|
| 8 |
# -----------------------------
|
| 9 |
st.set_page_config(page_title="Kurdish Translator", layout="wide")
|
| 10 |
|
| 11 |
+
# Initialize session state
|
| 12 |
+
if "theme" not in st.session_state:
|
| 13 |
+
st.session_state.theme = "light"
|
| 14 |
+
if "src_lang" not in st.session_state:
|
| 15 |
+
st.session_state.src_lang = "English"
|
| 16 |
+
if "tgt_lang" not in st.session_state:
|
| 17 |
+
st.session_state.tgt_lang = "Kurdish"
|
| 18 |
+
if "output_text" not in st.session_state:
|
| 19 |
+
st.session_state.output_text = ""
|
| 20 |
+
|
| 21 |
+
# Theme colors
|
| 22 |
+
THEME = st.session_state.theme
|
| 23 |
bg = "#111111" if THEME == "dark" else "#ffffff"
|
| 24 |
fg = "#ffffff" if THEME == "dark" else "#000000"
|
| 25 |
card_bg = "rgba(255,255,255,0.07)" if THEME == "dark" else "rgba(0,0,0,0.05)"
|
| 26 |
|
| 27 |
st.markdown(
|
| 28 |
+
f"""<style>
|
|
|
|
| 29 |
body {{
|
| 30 |
background-color: {bg};
|
| 31 |
color: {fg};
|
| 32 |
}}
|
|
|
|
| 33 |
.dots-loader {{
|
| 34 |
display: flex;
|
| 35 |
justify-content: center;
|
|
|
|
| 45 |
}}
|
| 46 |
.dots-loader div:nth-child(2) {{ animation-delay: 0.2s; }}
|
| 47 |
.dots-loader div:nth-child(3) {{ animation-delay: 0.4s; }}
|
|
|
|
| 48 |
@keyframes bounce {{
|
| 49 |
from {{ transform: translateY(0); }}
|
| 50 |
to {{ transform: translateY(-12px); }}
|
| 51 |
}}
|
|
|
|
| 52 |
.output-box {{
|
| 53 |
padding: 20px;
|
| 54 |
border-radius: 12px;
|
| 55 |
background: {card_bg};
|
| 56 |
font-size: 18px;
|
| 57 |
min-height: 120px;
|
| 58 |
+
word-wrap: break-word;
|
| 59 |
}}
|
| 60 |
+
</style>""",
|
|
|
|
| 61 |
unsafe_allow_html=True
|
| 62 |
)
|
| 63 |
|
| 64 |
# -----------------------------
|
| 65 |
+
# LOAD MODEL
|
| 66 |
# -----------------------------
|
| 67 |
BASE_MODEL = "facebook/nllb-200-distilled-600M"
|
| 68 |
LORA_REPO = "junaid17/nllb-kurdish-lora"
|
| 69 |
|
|
|
|
| 70 |
@st.cache_resource
|
| 71 |
def load_model():
|
| 72 |
tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)
|
|
|
|
| 74 |
model = PeftModel.from_pretrained(base, LORA_REPO)
|
| 75 |
return model, tokenizer
|
| 76 |
|
|
|
|
| 77 |
model, tokenizer = load_model()
|
| 78 |
|
|
|
|
| 79 |
# -----------------------------
|
| 80 |
# TRANSLATION FUNCTION
|
| 81 |
# -----------------------------
|
|
|
|
| 90 |
)
|
| 91 |
return translator(text)[0]["translation_text"]
|
| 92 |
|
| 93 |
+
# -----------------------------
|
| 94 |
+
# CALLBACK FUNCTIONS
|
| 95 |
+
# -----------------------------
|
| 96 |
+
def swap_languages():
|
| 97 |
+
st.session_state.src_lang, st.session_state.tgt_lang = (
|
| 98 |
+
st.session_state.tgt_lang,
|
| 99 |
+
st.session_state.src_lang
|
| 100 |
+
)
|
| 101 |
+
|
| 102 |
+
def toggle_theme():
|
| 103 |
+
st.session_state.theme = "dark" if st.session_state.theme == "light" else "light"
|
| 104 |
+
|
| 105 |
+
def copy_to_clipboard():
|
| 106 |
+
if st.session_state.output_text:
|
| 107 |
+
st.toast("β Text copied to clipboard!", icon="β
")
|
| 108 |
|
| 109 |
# -----------------------------
|
| 110 |
# UI
|
|
|
|
| 114 |
col1, colSwap, colTheme, col2 = st.columns([1, 0.6, 0.8, 1])
|
| 115 |
|
| 116 |
with col1:
|
| 117 |
+
src_lang = st.selectbox(
|
| 118 |
+
"From:",
|
| 119 |
+
["English", "Kurdish"],
|
| 120 |
+
key="src_lang"
|
| 121 |
+
)
|
| 122 |
|
| 123 |
with colSwap:
|
| 124 |
+
st.button("β Swap", use_container_width=True, on_click=swap_languages)
|
| 125 |
|
| 126 |
with colTheme:
|
| 127 |
+
theme_icon = "π" if THEME == "light" else "βοΈ"
|
| 128 |
+
st.button(theme_icon, use_container_width=True, on_click=toggle_theme)
|
| 129 |
|
| 130 |
with col2:
|
| 131 |
+
tgt_lang = st.selectbox(
|
| 132 |
+
"To:",
|
| 133 |
+
["Kurdish", "English"],
|
| 134 |
+
key="tgt_lang"
|
| 135 |
+
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 136 |
|
| 137 |
lang_codes = {
|
| 138 |
"English": "eng_Latn",
|
| 139 |
"Kurdish": "ckb_Arab"
|
| 140 |
}
|
| 141 |
|
| 142 |
+
src_code = lang_codes[st.session_state.src_lang]
|
| 143 |
+
tgt_code = lang_codes[st.session_state.tgt_lang]
|
| 144 |
|
| 145 |
text = st.text_area("Enter text:", height=180)
|
| 146 |
|
|
|
|
| 147 |
# -----------------------------
|
| 148 |
+
# TRANSLATE BUTTON
|
| 149 |
# -----------------------------
|
| 150 |
+
if st.button("Translate", type="primary", use_container_width=True):
|
| 151 |
if not text.strip():
|
| 152 |
st.warning("Please enter some text.")
|
| 153 |
else:
|
| 154 |
+
with st.spinner("Translating..."):
|
| 155 |
+
try:
|
| 156 |
+
output = translate(text, src_code, tgt_code)
|
| 157 |
+
st.session_state.output_text = output
|
| 158 |
+
except Exception as e:
|
| 159 |
+
st.error(f"Translation error: {str(e)}")
|
| 160 |
+
st.session_state.output_text = ""
|
| 161 |
+
|
| 162 |
+
# Display output if available
|
| 163 |
+
if st.session_state.output_text:
|
| 164 |
+
st.subheader("Output")
|
| 165 |
+
st.markdown(
|
| 166 |
+
f"<div class='output-box'>{st.session_state.output_text}</div>",
|
| 167 |
+
unsafe_allow_html=True
|
| 168 |
+
)
|
| 169 |
+
|
| 170 |
+
# Copy button with actual clipboard functionality
|
| 171 |
+
col_copy, col_empty = st.columns([1, 3])
|
| 172 |
+
with col_copy:
|
| 173 |
+
if st.button("π Copy Output", use_container_width=True):
|
| 174 |
+
# Use Streamlit's native clipboard through components
|
| 175 |
+
st.write(f"""
|
| 176 |
+
<script>
|
| 177 |
+
navigator.clipboard.writeText(`{st.session_state.output_text}`);
|
| 178 |
+
</script>
|
| 179 |
+
""", unsafe_allow_html=True)
|
| 180 |
+
st.toast("β Copied to clipboard!", icon="β
")
|