tanner/min-dalle-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

removed deepcopy, delete expendable parameters after use

Browse Source
This commit is contained in:
Brett Kuprel 2022-06-30 11:09:09 -04:00
parent 41a44068d0
commit b55bcba4c0
3 changed files with 13 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
min_dalle/load_params.py
View File

@ -82,11 +82,10 @@ def convert_dalle_bart_torch_from_flax_params(
layer_count: int, layer_count: int,
is_encoder: bool is_encoder: bool
) -> dict: ) -> dict:
P = deepcopy(params) P: Dict[str, numpy.ndarray] = flatten_dict(params, sep='.')
P: Dict[str, numpy.ndarray] = flatten_dict(P, sep='.')
for i in P: for i in P:
P[i] = torch.tensor(P[i]) P[i] = torch.tensor(P[i]).to(torch.float16)
for i in list(P): for i in list(P):
if 'kernel' in i: if 'kernel' in i:

10
min_dalle/min_dalle_flax.py
View File

@ -28,7 +28,7 @@ class MinDalleFlax(MinDalleBase):
text_token_count = self.config['max_text_length'], text_token_count = self.config['max_text_length'],
text_vocab_count = self.config['encoder_vocab_size'], text_vocab_count = self.config['encoder_vocab_size'],
layer_count = self.config['encoder_layers'] layer_count = self.config['encoder_layers']
).bind({'params': self.model_params['encoder']}) ).bind({'params': self.model_params.pop('encoder')})
def init_decoder(self): def init_decoder(self):
@ -53,13 +53,17 @@ class MinDalleFlax(MinDalleBase):
encoder_state = self.encoder(text_tokens) encoder_state = self.encoder(text_tokens)
if self.is_expendable: del self.encoder if self.is_expendable: del self.encoder
if self.is_expendable: self.init_decoder() if self.is_expendable:
self.init_decoder()
params = self.model_params.pop('decoder')
else:
params = self.model_params['decoder']
print("sampling image tokens") print("sampling image tokens")
image_tokens = self.decoder.sample_image_tokens( image_tokens = self.decoder.sample_image_tokens(
text_tokens, text_tokens,
encoder_state, encoder_state,
jax.random.PRNGKey(seed), jax.random.PRNGKey(seed),
self.model_params['decoder'] params
) )
if self.is_expendable: del self.decoder if self.is_expendable: del self.decoder

6
min_dalle/min_dalle_torch.py
View File

@ -40,12 +40,13 @@ class MinDalleTorch(MinDalleBase):
glu_embed_count = self.config['encoder_ffn_dim'] glu_embed_count = self.config['encoder_ffn_dim']
) )
params = convert_dalle_bart_torch_from_flax_params( params = convert_dalle_bart_torch_from_flax_params(
self.model_params['encoder'], self.model_params.pop('encoder'),
layer_count=self.config['encoder_layers'], layer_count=self.config['encoder_layers'],
is_encoder=True is_encoder=True
) )
self.encoder.load_state_dict(params, strict=False) self.encoder.load_state_dict(params, strict=False)
if torch.cuda.is_available(): self.encoder = self.encoder.cuda() if torch.cuda.is_available(): self.encoder = self.encoder.cuda()
del params
def init_decoder(self): def init_decoder(self):
@ -63,12 +64,13 @@ class MinDalleTorch(MinDalleBase):
is_verbose = True is_verbose = True
) )
params = convert_dalle_bart_torch_from_flax_params( params = convert_dalle_bart_torch_from_flax_params(
self.model_params['decoder'], self.model_params.pop('decoder'),
layer_count=self.config['decoder_layers'], layer_count=self.config['decoder_layers'],
is_encoder=False is_encoder=False
) )
self.decoder.load_state_dict(params, strict=False) self.decoder.load_state_dict(params, strict=False)
if torch.cuda.is_available(): self.decoder = self.decoder.cuda() if torch.cuda.is_available(): self.decoder = self.decoder.cuda()
del params
def init_detokenizer(self): def init_detokenizer(self):