---
title: "GET /training-info"
method: GET
path: "/training-info"
---

# GET /training-info

`GET /training-info`

Get info necessary to train a model

## Query parameters

- `fineTuneId` string, required

## Response `200`

Successful response

- object
  - `trainingDataUrl` string, required
  - `huggingFaceModelId` string, required
  - `trainingConfig` object, required
    - `model_type` 'LlamaForCausalLM' | 'MistralForCausalLM' | 'MixtralForCausalLM', required
    - `tokenizer_type` string, required
    - `is_llama_derived_model` boolean
    - `is_mistral_derived_model` boolean
    - `load_in_8bit` boolean, required
    - `load_in_4bit` boolean
    - `adapter` 'lora' | 'qlora', required
    - `model_config` object
      - `output_router_logits` boolean
    - `sequence_len` number, required
    - `sample_packing` boolean, required
    - `eval_sample_packing` boolean
    - `pad_to_sequence_len` boolean
    - `lora_r` number, required
    - `lora_alpha` number, required
    - `lora_dropout` number, required
    - `lora_target_linear` boolean, required
    - `lora_target_modules` string[]
    - `gradient_accumulation_steps` number, required
    - `micro_batch_size` number, required
    - `optimizer` string, required
    - `lr_scheduler` string, required
    - `learning_rate` number, required
    - `train_on_inputs` boolean, required
    - `group_by_length` boolean, required
    - `bf16` boolean, required
    - `fp16` boolean, required
    - `tf32` boolean, required
    - `gradient_checkpointing` boolean, required
    - `flash_attention` boolean, required
    - `warmup_steps` number, required
    - `weight_decay` number, required
    - `special_tokens` object, required
      - `bos_token` string, required
      - `eos_token` string, required
      - `unk_token` string, required
    - `base_model` string, required
    - `base_model_config` string
    - `datasets` object[], required
      - `path` string, required
      - `type` string, required
    - `dataset_processes` number
    - `val_set_size` number, required
    - `output_dir` string
    - `wandb_project` string
    - `wandb_run_id` string
    - `num_epochs` number, required
    - `logging_steps` number
    - `save_safetensors` boolean
    - `eval_steps` number
    - `strict` boolean
    - `save_strategy` string
    - `deepspeed` string
  - `fireworksBaseModel` string

## Other responses

- `default` — Error response

## Changes

- **2026-07-17** `1c16a82b7996` — 1 info
  - the response property `trainingConfig/base_model` became required for the status `200`
- **2024-02-15** `06ac3b58a5c3` — 1 info
  - the response property `trainingConfig/val_set_size` became required for the status `200`
- **2024-02-09** `0b81ef6594f4` — 1 warning, 4 info
  - added the new `MixtralForCausalLM` enum value to the `trainingConfig/model_type` response property for the response status `200`
  - added the optional property `fireworksBaseModel` to the response with the `200` status
  - added the optional property `trainingConfig/deepspeed` to the response with the `200` status
  - added the optional property `trainingConfig/eval_sample_packing` to the response with the `200` status
  - …1 more
- **2024-01-02** `0a168cd5ed42` — 44 info
  - added the optional property `trainingConfig/base_model` to the response with the `200` status
  - added the optional property `trainingConfig/base_model_config` to the response with the `200` status
  - added the optional property `trainingConfig/dataset_processes` to the response with the `200` status
  - added the optional property `trainingConfig/eval_steps` to the response with the `200` status
  - …40 more
- …earlier changes not shown

[Full history](https://skmtc.dev/openpipe/apis/openpipe-internal-api/changes/training-info/get.md)

---

[API](https://skmtc.dev/openpipe/apis/openpipe-internal-api.md) · [All operations](https://skmtc.dev/openpipe/apis/openpipe-internal-api/llms.txt) · [OpenAPI document](https://skmtc-service-production.skmtc.workers.dev/v1/apis/openpipe/openpipe-internal-api/revisions/1c16a82b7996/schema)