POST
/
models
/
v2
/
almanach
/
camembert-base
curl --request POST \
  --url https://api.bytez.com/models/v2/almanach/camembert-base \
  --header 'Authorization: <api-key>' \
  --header 'Content-Type: application/json' \
  --data '{
  "text": "<string>",
  "messages": [
    {
      "role": "system",
      "content": "<string>"
    }
  ],
  "url": "<string>",
  "base64": "<string>",
  "question": "<string>",
  "context": "<string>",
  "candidate_labels": [
    "<string>"
  ],
  "stream": true,
  "json": true,
  "params": {
    "min_length": 123,
    "max_length": 123,
    "temperature": 123
  }
}'
{
  "output": [
    {
      "score": 0.85,
      "token": 83,
      "token_str": "world",
      "sequence": "Hello world!"
    },
    {
      "score": 0.1,
      "token": 84,
      "token_str": "there",
      "sequence": "Hello there!"
    }
  ]
}

​
Fill Mask

The fill-mask task uses a masked language model to predict masked tokens in a given text.

​
Example

​
Request

curl --location 'https://api.bytez.com/models/v2/almanach/camembert-base' \
--header 'Authorization: Key your-api-key-here' \
--header 'Content-Type: application/json' \
--data '{
    "text": "Hello <mask>"
}'

Authorizations

​
Authorization
string
header
required

Set Authorization header to Key BYTEZ_KEY (e.g. 'Authorization': 'Key ABC123')

Body

application/json
Universal schema supporting text, chat, image, audio, video, QA, zero-shot, etc.
​
text
string

Input text for NLP tasks.

​
messages
object[]

Conversation history.

​
url
string

URL to image/audio/video.

​
base64
string

Base64-encoded image/audio/video.

​
question
string

Question text (QA tasks).

​
context
string

Context paragraph (QA tasks).

​
candidate_labels
string[]

Candidate labels (zero-shot).

​
stream
boolean

Enable text streaming.

​
json
boolean

Stream media instead of JSON.

​
params
object

Model-specific parameters.

Response

200 - application/json
Successful response with predictions for the masked token.
​
output
object[]