Instructions to use philipperen55/bloom-7b1-datasetSFT73_lora4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use philipperen55/bloom-7b1-datasetSFT73_lora4 with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("bigscience/bloom-7b1")
model = PeftModel.from_pretrained(base_model, "philipperen55/bloom-7b1-datasetSFT73_lora4")

Transformers

How to use philipperen55/bloom-7b1-datasetSFT73_lora4 with Transformers:

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

pipe = pipeline("text-generation", model="philipperen55/bloom-7b1-datasetSFT73_lora4")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("philipperen55/bloom-7b1-datasetSFT73_lora4")
model = AutoModelForCausalLM.from_pretrained("philipperen55/bloom-7b1-datasetSFT73_lora4", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use philipperen55/bloom-7b1-datasetSFT73_lora4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "philipperen55/bloom-7b1-datasetSFT73_lora4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "philipperen55/bloom-7b1-datasetSFT73_lora4",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/philipperen55/bloom-7b1-datasetSFT73_lora4

SGLang

How to use philipperen55/bloom-7b1-datasetSFT73_lora4 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 "philipperen55/bloom-7b1-datasetSFT73_lora4" \
    --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": "philipperen55/bloom-7b1-datasetSFT73_lora4",
		"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 "philipperen55/bloom-7b1-datasetSFT73_lora4" \
        --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": "philipperen55/bloom-7b1-datasetSFT73_lora4",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use philipperen55/bloom-7b1-datasetSFT73_lora4 with Docker Model Runner:
```
docker model run hf.co/philipperen55/bloom-7b1-datasetSFT73_lora4
```

See axolotl config

axolotl version: 0.17.0.dev0

base_model: bigscience/bloom-7b1
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code: true

load_in_8bit: false
load_in_4bit: false


datasets:
  - path: philipperen55/datasetSFT74bloom 
    ds_type: json
    data_files: datasetSFT74bloom.jsonl
    type: input_output
  

train_on_inputs: false
add_eos_token: false

  
dataset_prepared_path: /workspace/prepared_data
val_set_size: 0.05
output_dir: /workspace/output


sequence_len: 2048
sample_packing: false
pad_to_sequence_len: false
group_by_length: true





adapter: lora
lora_r: 32
lora_alpha: 64
lora_dropout: 0.1
lora_target_modules:
  - query_key_value
  - dense
  - dense_h_to_4h
  - dense_4h_to_h


gradient_accumulation_steps: 4
micro_batch_size: 4 
num_epochs: 3
learning_rate: 5e-5
lr_scheduler: cosine
warmup_ratio: 0.05
optimizer: adamw_torch_fused
weight_decay: 0.01
max_grad_norm: 1.0


bf16: true
fp16: false
tf32: true



overrides_of_model_config:
  use_cache: false
  

gradient_checkpointing: false


seed: 42
#mettre 24 si ya plus de 24 vspu, sinon mettre 16 si ya 24vcpu
dataset_num_proc: 16


logging_steps: 5
save_steps: 50
eval_strategy: steps 
eval_steps: 50
save_total_limit: 1


wandb_project: datasetSFT73


hub_model_id: philipperen55/bloom-7b1-datasetSFT73_lora4
push_to_hub: true
hub_strategy: every_save

bloom-7b1-datasetSFT73_lora4

This model is a fine-tuned version of bigscience/bloom-7b1 on the philipperen55/datasetSFT74bloom dataset. It achieves the following results on the evaluation set:

Loss: 2.2421
Ppl: 9.4130
Memory/max Active (gib): 103.37
Memory/max Allocated (gib): 103.37
Memory/device Reserved (gib): 162.69

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: 5e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Use OptimizerNames.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: 14
training_steps: 288

Training results

Training Loss	Epoch	Step	Validation Loss	Ppl	Active (gib)	Allocated (gib)	Reserved (gib)
No log	0	0	2.5977	13.4329	28.21	28.21	40.22
2.2493	0.5208	50	2.3535	10.5224	28.69	28.69	144.01
2.2127	1.0417	100	2.2872	9.8469	28.69	28.69	146.06
2.2381	1.5625	150	2.2652	9.6331	28.69	28.69	155.28
2.2350	2.0833	200	2.2565	9.5492	28.69	28.69	145.67
2.2595	2.6042	250	2.2430	9.4211	28.69	28.69	148.38
1.9304	3.0	288	2.2421	9.4130	103.37	103.37	162.69

Framework versions

PEFT 0.19.1
Transformers 5.9.0
Pytorch 2.10.0+cu128
Datasets 4.8.5
Tokenizers 0.22.2

Downloads last month: 42

Model tree for philipperen55/bloom-7b1-datasetSFT73_lora4

Base model

bigscience/bloom-7b1

Adapter

(39)

this model