Revert "[vicuna] Add streaming of tokens (#1587)" (#1588)

This reverts commit 4d55e51d46.
This commit is contained in:
powderluv
2023-06-23 10:29:00 -07:00
committed by GitHub
parent 4d55e51d46
commit 726d73d6ba
2 changed files with 5 additions and 6 deletions

View File

@@ -464,7 +464,6 @@ class Vicuna(SharkLLMBase):
logits = generated_token_op["logits"]
pkv = generated_token_op["pkv"]
detok = generated_token_op["detok"]
yield detok
res.append(detok)
res_tokens.append(token)
@@ -506,8 +505,6 @@ class Vicuna(SharkLLMBase):
res.append(detok)
if cli:
print(f"{detok}", end=" ", flush=True)
yield detok
if self.device == "cuda":
del sec_vic, pkv, logits
torch.cuda.empty_cache()

View File

@@ -65,14 +65,16 @@ def chat(curr_system_message, history, model, device, precision):
)
prompt = messages.strip()
print("prompt = ", prompt)
sentence = vicuna_model.generate(prompt)
partial_text = ""
for new_text in vicuna_model.generate(prompt):
for new_text in sentence.split(" "):
# print(new_text)
partial_text += new_text + " "
history[-1][1] = partial_text
# Yield an empty string to cleanup the message textbox and the updated conversation history
yield history
history[-1][1] = sentence
return history
# else Model is StableLM
@@ -122,13 +124,13 @@ with gr.Blocks(title="Chatbot") as stablelm_chat:
"TheBloke/vicuna-7B-1.1-HF",
],
)
supported_devices = available_devices + ["AMD-AIE"]
supported_devices = available_devices
enabled = len(supported_devices) > 0
device = gr.Dropdown(
label="Device",
value=supported_devices[0]
if enabled
else "No devices supported for now",
else "Only CUDA Supported for now",
choices=supported_devices,
interactive=enabled,
)