take lm mixed
Browse files- create_lm.ipynb +19 -19
- language_model/5gram.bin +2 -2
- language_model/unigrams.txt +0 -0
create_lm.ipynb
CHANGED
@@ -3,7 +3,7 @@
|
|
3 |
{
|
4 |
"cell_type": "code",
|
5 |
"execution_count": 6,
|
6 |
-
"id": "
|
7 |
"metadata": {},
|
8 |
"outputs": [],
|
9 |
"source": [
|
@@ -15,7 +15,7 @@
|
|
15 |
{
|
16 |
"cell_type": "code",
|
17 |
"execution_count": 11,
|
18 |
-
"id": "
|
19 |
"metadata": {},
|
20 |
"outputs": [],
|
21 |
"source": [
|
@@ -26,7 +26,7 @@
|
|
26 |
{
|
27 |
"cell_type": "code",
|
28 |
"execution_count": 4,
|
29 |
-
"id": "
|
30 |
"metadata": {},
|
31 |
"outputs": [
|
32 |
{
|
@@ -142,7 +142,7 @@
|
|
142 |
{
|
143 |
"cell_type": "code",
|
144 |
"execution_count": 12,
|
145 |
-
"id": "
|
146 |
"metadata": {},
|
147 |
"outputs": [],
|
148 |
"source": [
|
@@ -157,7 +157,7 @@
|
|
157 |
{
|
158 |
"cell_type": "code",
|
159 |
"execution_count": 13,
|
160 |
-
"id": "
|
161 |
"metadata": {},
|
162 |
"outputs": [
|
163 |
{
|
@@ -182,7 +182,7 @@
|
|
182 |
{
|
183 |
"cell_type": "code",
|
184 |
"execution_count": 14,
|
185 |
-
"id": "
|
186 |
"metadata": {},
|
187 |
"outputs": [
|
188 |
{
|
@@ -207,7 +207,7 @@
|
|
207 |
{
|
208 |
"cell_type": "code",
|
209 |
"execution_count": 15,
|
210 |
-
"id": "
|
211 |
"metadata": {},
|
212 |
"outputs": [],
|
213 |
"source": [
|
@@ -218,7 +218,7 @@
|
|
218 |
{
|
219 |
"cell_type": "code",
|
220 |
"execution_count": 7,
|
221 |
-
"id": "
|
222 |
"metadata": {},
|
223 |
"outputs": [],
|
224 |
"source": [
|
@@ -238,8 +238,8 @@
|
|
238 |
},
|
239 |
{
|
240 |
"cell_type": "code",
|
241 |
-
"execution_count":
|
242 |
-
"id": "
|
243 |
"metadata": {},
|
244 |
"outputs": [],
|
245 |
"source": [
|
@@ -250,8 +250,8 @@
|
|
250 |
},
|
251 |
{
|
252 |
"cell_type": "code",
|
253 |
-
"execution_count":
|
254 |
-
"id": "
|
255 |
"metadata": {},
|
256 |
"outputs": [],
|
257 |
"source": [
|
@@ -261,8 +261,8 @@
|
|
261 |
},
|
262 |
{
|
263 |
"cell_type": "code",
|
264 |
-
"execution_count":
|
265 |
-
"id": "
|
266 |
"metadata": {},
|
267 |
"outputs": [
|
268 |
{
|
@@ -287,8 +287,8 @@
|
|
287 |
},
|
288 |
{
|
289 |
"cell_type": "code",
|
290 |
-
"execution_count":
|
291 |
-
"id": "
|
292 |
"metadata": {},
|
293 |
"outputs": [],
|
294 |
"source": [
|
@@ -303,8 +303,8 @@
|
|
303 |
},
|
304 |
{
|
305 |
"cell_type": "code",
|
306 |
-
"execution_count":
|
307 |
-
"id": "
|
308 |
"metadata": {},
|
309 |
"outputs": [],
|
310 |
"source": [
|
@@ -314,7 +314,7 @@
|
|
314 |
{
|
315 |
"cell_type": "code",
|
316 |
"execution_count": null,
|
317 |
-
"id": "
|
318 |
"metadata": {},
|
319 |
"outputs": [],
|
320 |
"source": []
|
|
|
3 |
{
|
4 |
"cell_type": "code",
|
5 |
"execution_count": 6,
|
6 |
+
"id": "0f0b980b",
|
7 |
"metadata": {},
|
8 |
"outputs": [],
|
9 |
"source": [
|
|
|
15 |
{
|
16 |
"cell_type": "code",
|
17 |
"execution_count": 11,
|
18 |
+
"id": "55293a8b",
|
19 |
"metadata": {},
|
20 |
"outputs": [],
|
21 |
"source": [
|
|
|
26 |
{
|
27 |
"cell_type": "code",
|
28 |
"execution_count": 4,
|
29 |
+
"id": "3b5a735e",
|
30 |
"metadata": {},
|
31 |
"outputs": [
|
32 |
{
|
|
|
142 |
{
|
143 |
"cell_type": "code",
|
144 |
"execution_count": 12,
|
145 |
+
"id": "153d3a5e",
|
146 |
"metadata": {},
|
147 |
"outputs": [],
|
148 |
"source": [
|
|
|
157 |
{
|
158 |
"cell_type": "code",
|
159 |
"execution_count": 13,
|
160 |
+
"id": "1f83c4d9",
|
161 |
"metadata": {},
|
162 |
"outputs": [
|
163 |
{
|
|
|
182 |
{
|
183 |
"cell_type": "code",
|
184 |
"execution_count": 14,
|
185 |
+
"id": "5c14f88e",
|
186 |
"metadata": {},
|
187 |
"outputs": [
|
188 |
{
|
|
|
207 |
{
|
208 |
"cell_type": "code",
|
209 |
"execution_count": 15,
|
210 |
+
"id": "b2d7665a",
|
211 |
"metadata": {},
|
212 |
"outputs": [],
|
213 |
"source": [
|
|
|
218 |
{
|
219 |
"cell_type": "code",
|
220 |
"execution_count": 7,
|
221 |
+
"id": "959cb6fa",
|
222 |
"metadata": {},
|
223 |
"outputs": [],
|
224 |
"source": [
|
|
|
238 |
},
|
239 |
{
|
240 |
"cell_type": "code",
|
241 |
+
"execution_count": 4,
|
242 |
+
"id": "3fdae043",
|
243 |
"metadata": {},
|
244 |
"outputs": [],
|
245 |
"source": [
|
|
|
250 |
},
|
251 |
{
|
252 |
"cell_type": "code",
|
253 |
+
"execution_count": 5,
|
254 |
+
"id": "2c3c2b49",
|
255 |
"metadata": {},
|
256 |
"outputs": [],
|
257 |
"source": [
|
|
|
261 |
},
|
262 |
{
|
263 |
"cell_type": "code",
|
264 |
+
"execution_count": 6,
|
265 |
+
"id": "103034b6",
|
266 |
"metadata": {},
|
267 |
"outputs": [
|
268 |
{
|
|
|
287 |
},
|
288 |
{
|
289 |
"cell_type": "code",
|
290 |
+
"execution_count": 7,
|
291 |
+
"id": "40dd5ada",
|
292 |
"metadata": {},
|
293 |
"outputs": [],
|
294 |
"source": [
|
|
|
303 |
},
|
304 |
{
|
305 |
"cell_type": "code",
|
306 |
+
"execution_count": 8,
|
307 |
+
"id": "754cd832",
|
308 |
"metadata": {},
|
309 |
"outputs": [],
|
310 |
"source": [
|
|
|
314 |
{
|
315 |
"cell_type": "code",
|
316 |
"execution_count": null,
|
317 |
+
"id": "577d9617",
|
318 |
"metadata": {},
|
319 |
"outputs": [],
|
320 |
"source": []
|
language_model/5gram.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7cd870f7ccfef9e9a9b8f53f6b5f22c4c1331fd8b352bcb73d8556d67062df4c
|
3 |
+
size 3499369566
|
language_model/unigrams.txt
CHANGED
The diff for this file is too large to render.
See raw diff
|
|