Instructions to use Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain with Transformers:

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

pipe = pipeline("text-generation", model="Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain

SGLang

How to use Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain 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 "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain" \
    --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": "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain",
		"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 "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain" \
        --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": "Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain with Docker Model Runner:
```
docker model run hf.co/Nagase-Kotono/LLaVA-42dot_LLM-1.3B-pretrain
```

Nagase-Kotono commited on Mar 22, 2024

Commit

a424b3c

verified ·

1 Parent(s): 153ed66

03.22.15:27

Browse files

Files changed (9) hide show

checkpoint-2000/config.json +48 -0
checkpoint-2000/mm_projector.bin +3 -0
checkpoint-4000/config.json +48 -0
checkpoint-4000/mm_projector.bin +3 -0
checkpoint-6000/config.json +48 -0
checkpoint-6000/mm_projector.bin +3 -0
config.json +48 -0
mm_projector.bin +3 -0
trainer_state.json +0 -0

checkpoint-2000/config.json ADDED Viewed

	@@ -0,0 +1,48 @@

+{
+  "_name_or_path": "42dot/42dot_LLM-SFT-1.3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 50257,
+  "eos_token_id": 50256,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_aspect_ratio": "square",
+  "image_projector_type": "mlp2x_gelu",
+  "initializer_range": 0.01,
+  "intermediate_size": 5632,
+  "max_position_embeddings": 4096,
+  "mm_hidden_size": 1024,
+  "mm_image_tower": "openai/clip-vit-large-patch14-336",
+  "mm_projector_lr": null,
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_video_tower": null,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "model_type": "llava_llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 32,
+  "pad_token_id": 50259,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "GPT2TokenizerFast",
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.0",
+  "tune_mm_mlp_adapter": true,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "video_global_proj": false,
+  "video_projector_type": "linear",
+  "video_spatial_proj": false,
+  "video_temproal_proj": false,
+  "vocab_size": 50304
+}

checkpoint-2000/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:4f4b601c0fb2e65898a67a8a6b4feb93befcb66a3235075535d470e28018df17
+size 12592829

checkpoint-4000/config.json ADDED Viewed

	@@ -0,0 +1,48 @@

+{
+  "_name_or_path": "42dot/42dot_LLM-SFT-1.3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 50257,
+  "eos_token_id": 50256,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_aspect_ratio": "square",
+  "image_projector_type": "mlp2x_gelu",
+  "initializer_range": 0.01,
+  "intermediate_size": 5632,
+  "max_position_embeddings": 4096,
+  "mm_hidden_size": 1024,
+  "mm_image_tower": "openai/clip-vit-large-patch14-336",
+  "mm_projector_lr": null,
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_video_tower": null,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "model_type": "llava_llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 32,
+  "pad_token_id": 50259,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "GPT2TokenizerFast",
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.0",
+  "tune_mm_mlp_adapter": true,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "video_global_proj": false,
+  "video_projector_type": "linear",
+  "video_spatial_proj": false,
+  "video_temproal_proj": false,
+  "vocab_size": 50304
+}

checkpoint-4000/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:3c289b245f583932e318a094fd6e02c0a7531344f70f1cbc235e6a77650a72f5
+size 12592829

checkpoint-6000/config.json ADDED Viewed

	@@ -0,0 +1,48 @@

+{
+  "_name_or_path": "42dot/42dot_LLM-SFT-1.3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 50257,
+  "eos_token_id": 50256,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_aspect_ratio": "square",
+  "image_projector_type": "mlp2x_gelu",
+  "initializer_range": 0.01,
+  "intermediate_size": 5632,
+  "max_position_embeddings": 4096,
+  "mm_hidden_size": 1024,
+  "mm_image_tower": "openai/clip-vit-large-patch14-336",
+  "mm_projector_lr": null,
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_video_tower": null,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "model_type": "llava_llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 32,
+  "pad_token_id": 50259,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "GPT2TokenizerFast",
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.0",
+  "tune_mm_mlp_adapter": true,
+  "use_cache": false,
+  "use_mm_proj": true,
+  "video_global_proj": false,
+  "video_projector_type": "linear",
+  "video_spatial_proj": false,
+  "video_temproal_proj": false,
+  "vocab_size": 50304
+}

checkpoint-6000/mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:0187d20fb39a446c7b11fadaff3e062196c27ed1ea48bfe257db4ec5a40d7c98
+size 12592829

config.json ADDED Viewed

	@@ -0,0 +1,48 @@

+{
+  "_name_or_path": "42dot/42dot_LLM-SFT-1.3B",
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 50257,
+  "eos_token_id": 50256,
+  "freeze_mm_mlp_adapter": false,
+  "hidden_act": "silu",
+  "hidden_size": 2048,
+  "image_aspect_ratio": "square",
+  "image_projector_type": "mlp2x_gelu",
+  "initializer_range": 0.01,
+  "intermediate_size": 5632,
+  "max_position_embeddings": 4096,
+  "mm_hidden_size": 1024,
+  "mm_image_tower": "openai/clip-vit-large-patch14-336",
+  "mm_projector_lr": null,
+  "mm_use_im_patch_token": false,
+  "mm_use_im_start_end": false,
+  "mm_video_tower": null,
+  "mm_vision_select_feature": "patch",
+  "mm_vision_select_layer": -2,
+  "model_type": "llava_llama",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 24,
+  "num_key_value_heads": 32,
+  "pad_token_id": 50259,
+  "pretraining_tp": 1,
+  "rms_norm_eps": 1e-06,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "tie_word_embeddings": false,
+  "tokenizer_class": "GPT2TokenizerFast",
+  "tokenizer_padding_side": "right",
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.0",
+  "tune_mm_mlp_adapter": true,
+  "use_cache": true,
+  "use_mm_proj": true,
+  "video_global_proj": false,
+  "video_projector_type": "linear",
+  "video_spatial_proj": false,
+  "video_temproal_proj": false,
+  "vocab_size": 50304
+}

mm_projector.bin ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:a00d0f3aaecf58b5fb47a7f5dbb8b05667a1a0feafe3c60e4326ef6010e566e6
+size 12592829

trainer_state.json ADDED Viewed

The diff for this file is too large to render. See raw diff