Instructions to use anhnct/codec with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use anhnct/codec with Transformers:

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

pipe = pipeline("text-generation", model="anhnct/codec")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("anhnct/codec")
model = AutoModelForCausalLM.from_pretrained("anhnct/codec")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use anhnct/codec with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "anhnct/codec"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "anhnct/codec",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/anhnct/codec

SGLang

How to use anhnct/codec 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 "anhnct/codec" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "anhnct/codec",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "anhnct/codec" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "anhnct/codec",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use anhnct/codec with Docker Model Runner:
```
docker model run hf.co/anhnct/codec
```

See axolotl config

axolotl version: 0.8.0

base_model: /root/anhnct/Spark-TTS-finetune/extend_vocab_pretrained/LLM
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

trust_remote_code: true

strict: false

datasets:
  - path: .
    data_files: ["/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/elevenlab_dataset_3.jsonl", "/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/elevenlab_dataset_4.jsonl", "/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/elevenlab_dataset_reflex.jsonl", "/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/elevenlab_slow.jsonl", "/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/hf_song_ngu.jsonl", "/root/anhnct/Spark-TTS-finetune/PROMPTS/product_ft_data/LibriTTS.jsonl"]
    type: completion
    
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/Simp_22_1_2026


sequence_len: 2048
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 8
num_epochs: 10
optimizer: adamw_torch_fused
lr_scheduler: cosine
learning_rate: 0.0002

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

gradient_checkpointing: false
gradient_checkpointing_kwargs:
  use_reentrant: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 50
xformers_attention:
flash_attention: true

warmup_steps: 10
evals_per_epoch: 1
save_steps: 10000
save_total_limit: 100
debug:
deepspeed:
weight_decay: 0.0

outputs/Simp_22_1_2026

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

Loss: 5.3568

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: 8
eval_batch_size: 8
seed: 42
optimizer: Use adamw_torch_fused 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: 10
num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0005	1	5.6361
4.5777	1.0	2216	5.3235
4.5116	2.0	4432	5.3313
4.4611	3.0	6648	5.3390
4.4496	4.0	8864	5.3471
4.4141	5.0	11080	5.3521
4.4031	6.0	13296	5.3541
4.4174	7.0	15512	5.3562
4.4071	8.0	17728	5.3561
4.4179	9.0	19944	5.3567
4.3882	10.0	22160	5.3568

Framework versions

Transformers 4.50.3
Pytorch 2.6.0+cu124
Datasets 3.5.0
Tokenizers 0.21.4

Downloads last month: 6

Safetensors

Model size

0.5B params

Tensor type

BF16