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

Libraries

How to use momergul/flamingo_test with Transformers:

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

pipe = pipeline("text-generation", model="momergul/flamingo_test", trust_remote_code=True)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("momergul/flamingo_test", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("momergul/flamingo_test", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use momergul/flamingo_test with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/momergul/flamingo_test

SGLang

How to use momergul/flamingo_test 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 "momergul/flamingo_test" \
    --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": "momergul/flamingo_test",
		"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 "momergul/flamingo_test" \
        --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": "momergul/flamingo_test",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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

flamingo_test / utils.py

momergul

Upload utils.py with huggingface_hub

f61542a verified over 1 year ago

Raw History Blame Contribute Delete

1.23 kB

import torch

def exists(val):
    return val is not None

# for controlling freezing during training of flamingo

def set_module_requires_grad_(module, requires_grad):
    for param in module.parameters():
        param.requires_grad = requires_grad

def freeze_all_layers_(module):
    set_module_requires_grad_(module, False)

def unfreeze_all_layers_(module):
    set_module_requires_grad_(module, True)

def freeze_model_and_make_eval_(model):
    model.eval()
    freeze_all_layers_(model)

def _make_att_wd_mask(
        input_ids_shape: torch.Size, 
        dtype: torch.dtype, device: torch.device, 
        past_key_values_length: int = 0,
        att_wd_size: int = 0,
    ):
    bsz, tgt_len = input_ids_shape
    mask = torch.full((tgt_len, tgt_len), torch.finfo(dtype).min, device=device)
    mask_cond = torch.arange(mask.size(-1), device=device)
    mask.masked_fill_(
            mask_cond > (mask_cond - att_wd_size).view(mask.size(-1), 1), 0)
    mask = mask.to(dtype)

    if past_key_values_length > 0:
        mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
    return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)

	import torch

	def exists(val):
	return val is not None

	# for controlling freezing during training of flamingo

	def set_module_requires_grad_(module, requires_grad):
	for param in module.parameters():
	param.requires_grad = requires_grad

	def freeze_all_layers_(module):
	set_module_requires_grad_(module, False)

	def unfreeze_all_layers_(module):
	set_module_requires_grad_(module, True)

	def freeze_model_and_make_eval_(model):
	model.eval()
	freeze_all_layers_(model)

	def _make_att_wd_mask(
	input_ids_shape: torch.Size,
	dtype: torch.dtype, device: torch.device,
	past_key_values_length: int = 0,
	att_wd_size: int = 0,
	):
	bsz, tgt_len = input_ids_shape
	mask = torch.full((tgt_len, tgt_len), torch.finfo(dtype).min, device=device)
	mask_cond = torch.arange(mask.size(-1), device=device)
	mask.masked_fill_(
	mask_cond > (mask_cond - att_wd_size).view(mask.size(-1), 1), 0)
	mask = mask.to(dtype)

	if past_key_values_length > 0:
	mask = torch.cat([torch.zeros(tgt_len, past_key_values_length, dtype=dtype, device=device), mask], dim=-1)
	return mask[None, None, :, :].expand(bsz, 1, tgt_len, tgt_len + past_key_values_length)