[Feature]: support LlavaForConditionalGeneration with turbomind inference

docs/en/multi_modal/llava_qwen.md

@@ -0,0 +1,172 @@
+# Llava-Qwen2
+
+LMDeploy supports the following llava-qwen2 series of models, which are detailed in the table below:
+
+|            Model            | Size | Supported Inference Engine |
+| :-------------------------: | :--: | :------------------------: |
+| Llava-interleave-qwen-7b-hf |  7B  |         TurboMind          |
+
+The next chapter demonstrates how to deploy an LlavaQwen2 model using LMDeploy, with [LlavaQwen2](https://huggingface.co/llava-hf/llava-interleave-qwen-7b-hf) as an example.
+
+## Installation
+
+Please install LMDeploy by following the [installation guide](../get_started/installation.md).
+
+Or, you can build a docker image to set up the inference environment. If the CUDA version on your host machine is `>=12.4`, you can run:
+
+```
+git clone https://github.com/InternLM/lmdeploy.git
+cd lmdeploy
+docker build --build-arg CUDA_VERSION=cu12 -t openmmlab/lmdeploy:llava_qwen2 . -f ./docker/Dockerfile
+```
+
+Otherwise, you can go with:
+
+```shell
+docker build --build-arg CUDA_VERSION=cu11 -t openmmlab/lmdeploy:llava_qwen2 . -f ./docker/Dockerfile
+```
+
+## Offline inference
+
+The following sample code shows the basic usage of VLM pipeline. For detailed information, please refer to [VLM Offline Inference Pipeline](./vl_pipeline.md)
+
+```python
+from lmdeploy import GenerationConfig, TurbomindEngineConfig, pipeline
+from lmdeploy.vl import load_image
+
+
+pipe = pipeline("llava-hf/llava-interleave-qwen-7b-hf", backend_config=TurbomindEngineConfig(cache_max_entry_count=0.5),
+    gen_config=GenerationConfig(max_new_tokens=512))
+
+image = load_image('https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-VL/assets/demo.jpeg')
+prompt = 'Describe the image.'
+print(f'prompt:{prompt}')
+response = pipe((prompt, image))
+print(response)
+
+```
+
+More examples are listed below:
+
+<details>
+  <summary>
+    <b>multi-image multi-round conversation, combined images</b>
+  </summary>
+
+```python
+from lmdeploy import pipeline, GenerationConfig
+
+pipe = pipeline('llava-hf/llava-interleave-qwen-7b-hf', log_level='INFO')
+messages = [
+    dict(role='user', content=[
+        dict(type='text', text='Describe the two images in detail.'),
+        dict(type='image_url', image_url=dict(url='https://raw.githubusercontent.com/QwenLM/Qwen-VL/master/assets/mm_tutorial/Beijing_Small.jpeg')),
+        dict(type='image_url', image_url=dict(url='https://raw.githubusercontent.com/QwenLM/Qwen-VL/master/assets/mm_tutorial/Chongqing_Small.jpeg'))
+    ])
+]
+out = pipe(messages, gen_config=GenerationConfig(top_k=1))
+
+messages.append(dict(role='assistant', content=out.text))
+messages.append(dict(role='user', content='What are the similarities and differences between these two images.'))
+out = pipe(messages, gen_config=GenerationConfig(top_k=1))
+```
+
+</details>
+
+<details>
+  <summary>
+    <b>image resolution for performance boost</b>
+  </summary>
+
+```python
+from lmdeploy import pipeline, GenerationConfig
+
+pipe = pipeline('llava-hf/llava-interleave-qwen-7b-hf', log_level='INFO')
+
+min_pixels = 64 * 28 * 28
+max_pixels = 64 * 28 * 28
+messages = [
+    dict(role='user', content=[
+        dict(type='text', text='Describe the two images in detail.'),
+        dict(type='image_url', image_url=dict(min_pixels=min_pixels, max_pixels=max_pixels, url='https://raw.githubusercontent.com/QwenLM/Qwen-VL/master/assets/mm_tutorial/Beijing_Small.jpeg')),
+        dict(type='image_url', image_url=dict(min_pixels=min_pixels, max_pixels=max_pixels, url='https://raw.githubusercontent.com/QwenLM/Qwen-VL/master/assets/mm_tutorial/Chongqing_Small.jpeg'))
+    ])
+]
+out = pipe(messages, gen_config=GenerationConfig(top_k=1))
+
+messages.append(dict(role='assistant', content=out.text))
+messages.append(dict(role='user', content='What are the similarities and differences between these two images.'))
+out = pipe(messages, gen_config=GenerationConfig(top_k=1))
+```
+
+</details>
+
+## Online serving
+
+You can launch the server by the `lmdeploy serve api_server` CLI:
+
+```shell
+lmdeploy serve api_server llava-hf/llava-interleave-qwen-7b-hf
+```
+
+You can also start the service using the aforementioned built docker image:
+
+```shell
+docker run --runtime nvidia --gpus all \
+    -v ~/.cache/huggingface:/root/.cache/huggingface \
+    --env "HUGGING_FACE_HUB_TOKEN=<secret>" \
+    -p 23333:23333 \
+    --ipc=host \
+    openmmlab/lmdeploy:llava_qwen2 \
+    lmdeploy serve api_server llava-hf/llava-interleave-qwen-7b-hf
+```
+
+The docker compose is another option. Create a `docker-compose.yml` configuration file in the root directory of the lmdeploy project as follows:
+
+```yaml
+version: '3.5'
+
+services:
+  lmdeploy:
+    container_name: lmdeploy
+    image: openmmlab/lmdeploy:llava_qwen2
+    ports:
+      - "23333:23333"
+    environment:
+      HUGGING_FACE_HUB_TOKEN: <secret>
+    volumes:
+      - ~/.cache/huggingface:/root/.cache/huggingface
+    stdin_open: true
+    tty: true
+    ipc: host
+    command: lmdeploy serve api_server llava-hf/llava-interleave-qwen-7b-hf
+    deploy:
+      resources:
+        reservations:
+          devices:
+            - driver: nvidia
+              count: "all"
+              capabilities: [gpu]
+```
+
+Then, you can execute the startup command as below:
+
+```shell
+docker-compose up -d
+```
+
+If you find the following logs after running `docker logs -f lmdeploy`, it means the service launches successfully.
+
+```text
+HINT:    Please open  http://0.0.0.0:23333   in a browser for detailed api usage!!!
+HINT:    Please open  http://0.0.0.0:23333   in a browser for detailed api usage!!!
+HINT:    Please open  http://0.0.0.0:23333   in a browser for detailed api usage!!!
+INFO:     Started server process [2439]
+INFO:     Waiting for application startup.
+INFO:     Application startup complete.
+INFO:     Uvicorn running on  http://0.0.0.0:23333  (Press CTRL+C to quit)
+```
+
+The arguments of `lmdeploy serve api_server` can be reviewed in detail by `lmdeploy serve api_server -h`.
+
+More information about `api_server` as well as how to access the service can be found from [here](api_server_vl.md)

lmdeploy/turbomind/deploy/source_model/__init__.py

@@ -5,6 +5,7 @@
 from .internlm2 import InternLM2Model  # noqa: F401
 from .internvl import InternVLModel  # noqa: F401
 from .llama import LlamaModel  # noqa: F401
+from .llava_qwen2 import LlavaQwen2Model  # noqa: F401
 from .meta_llama import MetaLlamaModel  # noqa: F401
 from .minicpmv import MiniCPMVModel  # noqa: F401
 from .mixtral import MixtralModel  # noqa: F401

lmdeploy/turbomind/deploy/source_model/llava_qwen2.py

@@ -0,0 +1,77 @@
+# Copyright (c) OpenMMLab. All rights reserved.
+import json
+import os.path as osp
+
+from .base import INPUT_MODELS
+from .llama import LlamaModel, LlamaReader
+
+
+class LlavaQwen2Reader(LlamaReader):
+    """LlavaQwen2Reader for llama model."""
+
+    attn_layer_prefix = 'language_model.model.layers'
+    attn_layer_patten = r'language_model.model.layers.([0-9]+).'
+    tok_embeddings_key = 'language_model.model.embed_tokens.weight'
+    norm_weight_key = 'language_model.model.norm.weight'
+    output_weight_key = 'language_model.lm_head.weight'
+
+    def __init__(self, new_params: dict, unused_params: dict, last_bin: bool,
+                 model_cfg: dict, policy):
+        model_cfg = model_cfg.get('text_config')
+        super().__init__(new_params, unused_params, last_bin, model_cfg,
+                         policy)
+
+
+@INPUT_MODELS.register_module(name='llava_qwen2')
+class LlavaQwen2Model(LlamaModel):
+    """LlavaQwen2Model model in hf format."""
+
+    def __init__(self, model_path: str, tokenizer_path: str, **kwargs):
+        super().__init__(model_path, tokenizer_path, **kwargs)
+        self.Reader = LlavaQwen2Reader
+
+    def model_info(self):
+        """Read model info."""
+        params_path = osp.join(self.model_path, 'config.json')
+        with open(params_path) as f:
+            model_arg = json.load(f)['text_config']
+            num_layer = model_arg.get('num_hidden_layers', 32)
+            norm_eps = model_arg.get('rms_norm_eps', 1e-6)
+            attn_head_num = model_arg.get('num_attention_heads', 32)
+            if 'num_key_value_heads' in model_arg:
+                kv_head_num = model_arg.get('num_key_value_heads', 32)
+            else:
+                kv_head_num = model_arg.get('num_attention_heads', 32)
+            rope_theta = float(model_arg.get('rope_theta', 10000.0))
+            max_position_embeddings = int(
+                model_arg.get('max_position_embeddings', 0))
+            rope_scaling = model_arg.get('rope_scaling', None)
+            scaling_factor = 0.0
+            use_dynamic_ntk = 0
+
+            # special for the model: llava-hf/llava-interleave-qwen-7b-hf
+            hidden_units = model_arg.get('hidden_size', 4096)
+            vocab_size = model_arg.get('vocab_size', 152000)
+            intermediate_size = model_arg.get('intermediate_size', 11008)
+            attn_bias = int(model_arg.get('attn_bias', 1))
+            use_logn_attn = int(model_arg.get('use_logn_attn', 0))
+
+            if isinstance(rope_scaling, dict):
+                scaling_type = model_arg['rope_scaling'].get('type', '')
+                scaling_factor = model_arg['rope_scaling'].get('factor', '')
+                if scaling_type == 'dynamic':
+                    use_dynamic_ntk = 1
+
+        return dict(num_layer=num_layer,
+                    norm_eps=norm_eps,
+                    head_num=attn_head_num,
+                    hidden_units=hidden_units,
+                    kv_head_num=kv_head_num,
+                    rope_theta=rope_theta,
+                    max_position_embeddings=max_position_embeddings,
+                    use_dynamic_ntk=use_dynamic_ntk,
+                    rope_scaling_factor=scaling_factor,
+                    inter_size=intermediate_size,
+                    use_logn_attn=use_logn_attn,
+                    attn_bias=attn_bias,
+                    vocab_size=vocab_size)

lmdeploy/turbomind/generate_gemm_config.py

@@ -54,10 +54,23 @@ def main(head_num: int = 32,
             from transformers import AutoConfig
             config = AutoConfig.from_pretrained(model_path,
                                                 trust_remote_code=True)
-            head_num = config.num_attention_heads
-            size_per_head = config.hidden_size // head_num
-            inter_size = config.intermediate_size
-            vocab_size = config.vocab_size
+            try:
+                head_num = config.num_attention_heads
+                size_per_head = config.hidden_size // head_num
+                inter_size = config.intermediate_size
+                vocab_size = config.vocab_size
+            except AttributeError as e:
+                if hasattr(config, 'text_config'):
+                    config = config.text_config
+                elif hasattr(config, 'llm_config'):
+                    config = config.llm_config
+                else:
+                    raise AttributeError(f'not found attribute in {config},\
+                            please check your model config file.{e}')
+                head_num = config.num_attention_heads
+                size_per_head = config.hidden_size // head_num
+                inter_size = config.intermediate_size
+                vocab_size = config.vocab_size
     for bsz in range(1, max_batch_size + 1):
         subprocess.call(
             f'{get_llama_gemm()} {bsz} 1 1 {head_num} {size_per_head}'

lmdeploy/turbomind/supported_models.py

@@ -25,6 +25,8 @@
     # llava
     LlavaLlamaForCausalLM='llama',
     LlavaMistralForCausalLM='llama',
+    # Llava_interleave
+    LlavaForConditionalGeneration='llava_qwen2',
     # xcomposer2
     InternLMXComposer2ForCausalLM='xcomposer2',
     # internvl
@@ -99,5 +101,7 @@ def _is_head_dim_128(cfg):
             elif arch == 'InternVLChatModel':
                 # internvl2-4b,internlm2-1b are not working yet
                 support_by_turbomind = _is_head_dim_128(cfg.llm_config)
+            elif arch == 'LlavaForConditionalGeneration':
+                support_by_turbomind = _is_head_dim_128(cfg.text_config)
 
     return support_by_turbomind