Instructions to use ptllama/llama3b-2e4-2m-test2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ptllama/llama3b-2e4-2m-test2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="ptllama/llama3b-2e4-2m-test2")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("ptllama/llama3b-2e4-2m-test2")
model = AutoModelForCausalLM.from_pretrained("ptllama/llama3b-2e4-2m-test2")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ptllama/llama3b-2e4-2m-test2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ptllama/llama3b-2e4-2m-test2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ptllama/llama3b-2e4-2m-test2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ptllama/llama3b-2e4-2m-test2

SGLang

How to use ptllama/llama3b-2e4-2m-test2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "ptllama/llama3b-2e4-2m-test2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ptllama/llama3b-2e4-2m-test2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "ptllama/llama3b-2e4-2m-test2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ptllama/llama3b-2e4-2m-test2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ptllama/llama3b-2e4-2m-test2 with Docker Model Runner:
```
docker model run hf.co/ptllama/llama3b-2e4-2m-test2
```

See axolotl config

axolotl version: 0.7.0

base_model: meta-llama/Llama-3.2-3B
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: ptllama/acemath_test_2
    type: completion

# pretraining_dataset:
#   - name:
#     path: ptllama/acemath_test
#     split:
#     text_column: text # column in dataset with the data, usually `text`
#     type: pretrain
#     trust_remote_code:
#     skip: # number of rows of data to skip over from the beginning

dataset_prepared_path: last_run_prepared
val_set_size: 0.01
output_dir: ./outputs/out

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true
eval_sample_packing: false

wandb_project: axolotl-pretraining
wandb_entity:
wandb_watch:
wandb_name: test2-2e4
wandb_log_model:

gradient_accumulation_steps: 16
micro_batch_size: 4
num_epochs: 1
optimizer: paged_adamw_8bit
lr_scheduler: cosine
learning_rate: 2e-4

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_ratio: 0.01
cosine_min_lr_ratio: 0.1
cosine_constant_lr_ratio: 0.9
evals_per_epoch: 2
eval_table_size:
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
  pad_token: <|end_of_text|>

outputs/out

This model is a fine-tuned version of meta-llama/Llama-3.2-3B on the ptllama/acemath_test_2 dataset. It achieves the following results on the evaluation set:

Loss: 0.2925

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0002
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 8
gradient_accumulation_steps: 16
total_train_batch_size: 512
total_eval_batch_size: 32
optimizer: Use paged_adamw_8bit with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 17
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.9006	0.0006	1	0.9558
0.2852	0.5001	862	0.2925

Framework versions

Transformers 4.48.3
Pytorch 2.4.0
Datasets 3.2.0
Tokenizers 0.21.0

Downloads last month: 3

Safetensors

Model size

3B params

Tensor type

BF16

Model tree for ptllama/llama3b-2e4-2m-test2

Base model

meta-llama/Llama-3.2-3B

Finetuned

(444)

this model

ptllama
/

llama3b-2e4-2m-test2

outputs/out

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Model tree for ptllama/llama3b-2e4-2m-test2

Dataset used to train ptllama/llama3b-2e4-2m-test2