Unclear instructions
#3
by
agoudarzi
- opened
I deployed this model on the Vertex AI but I get Incomplete generation error:
prompt = """
### USER:\n Apa sentimen dari kalimat berikut ini?
Kalimat: Buku ini sangat membosankan.
Jawaban: \n\n### RESPONSE:\n
"""
messages = { "instances" : [ { "inputs" : prompt}],
"parameters" : { "max_new_tokens" : 100} }
headers = {
"Content-Type": "application/json",
"Authorization" : f"Bearer {token}"
}
print( headers)
url = f"https://us-central1-aiplatform.googleapis.com/v1/projects/{PROJECT_ID}/locations/us-central1/endpoints/{ENDPOINT_ID}:predict"
print(url)
t0 = time.time()
results = requests.post(url, json=messages, headers=headers)
t1 = time.time()
print(results.content)
print(t1-t0)
Dear @agoudarzi ,
Thank you for your interests in SEA-LION.
Could you elaborate more on the error you have encountered by sharing any error messages you have received when running this code?
Or do you meant the generated text was incomplete? If yes, could you kindly increase the value for max_new_tokens to a higher value (e.g. 1000) and see if the model is able to completely answer your query?
Thank you
Raymond