feat: Support for batch embedding

[hiro-v]

Problem

• I am trying to integrate Jan RAG feature using Nitro embedding at POST http://localhost:3928/v1/embeddings
• It supports the input in payload as string by default, however as the application requires batch embedding, nitro yields 500

Success Criteria

• Load model with embedding true

curl --location 'http://127.0.0.1:3928/inferences/llamacpp/loadmodel' \
--header 'Content-Type: application/json' \
--data '{
   "llama_model_path": "/Users/hiro/Downloads/ggml-model-q4_k.gguf",
   "ctx_len": 2048,
   "ngl": 100,
   "cont_batching": false,
   "embedding": true,
   "system_prompt": "",
   "user_prompt": "\n### Instruction:\n",
   "ai_prompt": "\n### Response:\n"
 }'

• Try to call the embedding endpoint with input as array

curl --location 'http://localhost:3928/v1/embeddings' \
--header 'Content-Type: application/json' \
--header 'Accept: text/event-stream' \
--header 'Access-Control-Allow-Origin: *' \
--data '{
    "input": ["Hello", "Nam", "here"],
    "model": "embedding",
    "encoding_format": "float"
}'

• The server returns 500

Additional context

• It's OAI compatible API: https://platform.openai.com/docs/api-reference/embeddings/create