Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -256,15 +256,17 @@ def generate_30(
|
|
| 256 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 257 |
print('catted shape: ', prompt_embeds.shape)
|
| 258 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
|
|
|
|
|
|
| 259 |
# 4. (Optional) Average the pooled embeddings
|
| 260 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 261 |
print('averaged shape: ', prompt_embeds.shape)
|
| 262 |
|
| 263 |
# 3. Concatenate the text_encoder_2 embeddings
|
| 264 |
-
prompt_embeds2 = torch.cat([prompt_embeds_a2, prompt_embeds_b2])
|
| 265 |
-
print('catted shape2: ', prompt_embeds.shape)
|
| 266 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
|
|
|
| 267 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
|
|
|
| 268 |
# 4. (Optional) Average the pooled embeddings
|
| 269 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 270 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|
|
@@ -379,15 +381,17 @@ def generate_60(
|
|
| 379 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 380 |
print('catted shape: ', prompt_embeds.shape)
|
| 381 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
|
|
|
|
|
|
| 382 |
# 4. (Optional) Average the pooled embeddings
|
| 383 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 384 |
print('averaged shape: ', prompt_embeds.shape)
|
| 385 |
|
| 386 |
# 3. Concatenate the text_encoder_2 embeddings
|
| 387 |
-
prompt_embeds2 = torch.cat([prompt_embeds_a2, prompt_embeds_b2])
|
| 388 |
-
print('catted shape2: ', prompt_embeds.shape)
|
| 389 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
|
|
|
| 390 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
|
|
|
| 391 |
# 4. (Optional) Average the pooled embeddings
|
| 392 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 393 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|
|
@@ -502,15 +506,17 @@ def generate_90(
|
|
| 502 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 503 |
print('catted shape: ', prompt_embeds.shape)
|
| 504 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
|
|
|
|
|
|
| 505 |
# 4. (Optional) Average the pooled embeddings
|
| 506 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 507 |
print('averaged shape: ', prompt_embeds.shape)
|
| 508 |
|
| 509 |
# 3. Concatenate the text_encoder_2 embeddings
|
| 510 |
-
prompt_embeds2 = torch.cat([prompt_embeds_a2, prompt_embeds_b2])
|
| 511 |
-
print('catted shape2: ', prompt_embeds.shape)
|
| 512 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
|
|
|
| 513 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
|
|
|
| 514 |
# 4. (Optional) Average the pooled embeddings
|
| 515 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 516 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|
|
|
|
| 256 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 257 |
print('catted shape: ', prompt_embeds.shape)
|
| 258 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
| 259 |
+
print('pooled shape: ', prompt_embeds.shape)
|
| 260 |
+
|
| 261 |
# 4. (Optional) Average the pooled embeddings
|
| 262 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 263 |
print('averaged shape: ', prompt_embeds.shape)
|
| 264 |
|
| 265 |
# 3. Concatenate the text_encoder_2 embeddings
|
|
|
|
|
|
|
| 266 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
| 267 |
+
print('catted pooled shape: ', prompt_embeds.shape)
|
| 268 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
| 269 |
+
print('catted combined pooled shape: ', prompt_embeds.shape)
|
| 270 |
# 4. (Optional) Average the pooled embeddings
|
| 271 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 272 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|
|
|
|
| 381 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 382 |
print('catted shape: ', prompt_embeds.shape)
|
| 383 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
| 384 |
+
print('pooled shape: ', prompt_embeds.shape)
|
| 385 |
+
|
| 386 |
# 4. (Optional) Average the pooled embeddings
|
| 387 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 388 |
print('averaged shape: ', prompt_embeds.shape)
|
| 389 |
|
| 390 |
# 3. Concatenate the text_encoder_2 embeddings
|
|
|
|
|
|
|
| 391 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
| 392 |
+
print('catted pooled shape: ', prompt_embeds.shape)
|
| 393 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
| 394 |
+
print('catted combined pooled shape: ', prompt_embeds.shape)
|
| 395 |
# 4. (Optional) Average the pooled embeddings
|
| 396 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 397 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|
|
|
|
| 506 |
prompt_embeds = torch.cat([prompt_embeds_a, prompt_embeds_b])
|
| 507 |
print('catted shape: ', prompt_embeds.shape)
|
| 508 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds_a, pooled_prompt_embeds_b])
|
| 509 |
+
print('pooled shape: ', prompt_embeds.shape)
|
| 510 |
+
|
| 511 |
# 4. (Optional) Average the pooled embeddings
|
| 512 |
prompt_embeds = torch.mean(prompt_embeds,dim=0,keepdim=True)
|
| 513 |
print('averaged shape: ', prompt_embeds.shape)
|
| 514 |
|
| 515 |
# 3. Concatenate the text_encoder_2 embeddings
|
|
|
|
|
|
|
| 516 |
pooled_prompt_embeds2 = torch.cat([pooled_prompt_embeds_a2, pooled_prompt_embeds_b2])
|
| 517 |
+
print('catted pooled shape: ', prompt_embeds.shape)
|
| 518 |
pooled_prompt_embeds = torch.cat([pooled_prompt_embeds, pooled_prompt_embeds2])
|
| 519 |
+
print('catted combined pooled shape: ', prompt_embeds.shape)
|
| 520 |
# 4. (Optional) Average the pooled embeddings
|
| 521 |
pooled_prompt_embeds = torch.mean(pooled_prompt_embeds,dim=0)
|
| 522 |
print('pooled averaged shape: ', pooled_prompt_embeds.shape)
|