Image-Text-to-Text
Transformers
Safetensors
mage_vl
multimodal
vision-language-model
mage-vl
video-understanding
streaming
conversational
custom_code
Instructions to use microsoft/Mage-VL with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use microsoft/Mage-VL with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="microsoft/Mage-VL", trust_remote_code=True) messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# Load model directly from transformers import AutoModelForImageTextToText model = AutoModelForImageTextToText.from_pretrained("microsoft/Mage-VL", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use microsoft/Mage-VL with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "microsoft/Mage-VL" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "microsoft/Mage-VL", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/microsoft/Mage-VL
- SGLang
How to use microsoft/Mage-VL 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 "microsoft/Mage-VL" \ --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": "microsoft/Mage-VL", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'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 "microsoft/Mage-VL" \ --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": "microsoft/Mage-VL", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Docker Model Runner
How to use microsoft/Mage-VL with Docker Model Runner:
docker model run hf.co/microsoft/Mage-VL
File size: 6,986 Bytes
12acbba | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 | /* Copyright 2020 InterDigital Communications, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#pragma once
#include <condition_variable>
#include <list>
#include <thread>
#include <vector>
#ifdef __GNUC__
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wpedantic"
#pragma GCC diagnostic ignored "-Wsign-compare"
#endif
#ifdef _MSC_VER
#pragma warning(disable : 4244)
#endif
#include "rans_byte.h"
#ifdef _MSC_VER
#pragma warning(default : 4244)
#endif
#ifdef __GNUC__
#pragma GCC diagnostic pop
#endif
#ifdef _MSC_VER
#define FORCE_INLINE __forceinline
#endif
#ifdef __GNUC__
#define FORCE_INLINE __attribute__((always_inline)) inline
#endif
struct RansSymbol {
uint16_t start;
uint16_t range; // range for normal coding and 0 for bypass coding
};
enum class WorkType {
EncodeDecodeY,
EncodeDecodeZ,
Flush,
};
struct PendingTask {
WorkType workType;
std::shared_ptr<std::vector<int16_t>> symbols_y;
std::shared_ptr<std::vector<int8_t>> symbols_z;
std::shared_ptr<std::vector<uint8_t>> indexes;
int total_size{ 0 };
int cdf_group_index{ 0 };
int start_offset{ 0 };
int per_channel_size{ 0 };
};
/* NOTE: Warning, we buffer everything for now... In case of large files we
* should split the bitstream into chunks... Or for a memory-bounded encoder
**/
class RansEncoderLib {
public:
RansEncoderLib();
virtual ~RansEncoderLib() = default;
RansEncoderLib(const RansEncoderLib&) = delete;
RansEncoderLib(RansEncoderLib&&) = delete;
RansEncoderLib& operator=(const RansEncoderLib&) = delete;
RansEncoderLib& operator=(RansEncoderLib&&) = delete;
void encode_y(const std::shared_ptr<std::vector<int16_t>> symbols, const int cdf_group_index);
void encode_z(const std::shared_ptr<std::vector<int8_t>> symbols, const int cdf_group_index,
const int start_offset, const int per_channel_size);
FORCE_INLINE void encode_y_internal(uint8_t*& ptr, RansState& rans,
const std::shared_ptr<std::vector<int16_t>> symbols,
const int cdf_group_index);
FORCE_INLINE void encode_z_internal(uint8_t*& ptr, RansState& rans,
const std::shared_ptr<std::vector<int8_t>> symbols,
const int cdf_group_index, const int start_offset,
const int per_channel_size);
FORCE_INLINE void encode_one_symbol(uint8_t*& ptr, RansState& rans, const int32_t symbol,
const int32_t cdf_size, const int32_t offset,
const std::vector<RansSymbol>& ransSymbols);
virtual void flush();
virtual std::shared_ptr<std::vector<uint8_t>> get_encoded_stream();
virtual void reset();
virtual int add_cdf(const std::shared_ptr<std::vector<std::vector<int32_t>>> cdfs,
const std::shared_ptr<std::vector<int32_t>> cdfs_sizes,
const std::shared_ptr<std::vector<int32_t>> offsets);
virtual void empty_cdf_buffer();
private:
std::shared_ptr<std::vector<uint8_t>> _stream;
std::vector<std::shared_ptr<std::vector<std::vector<RansSymbol>>>> _ransSymbols;
std::vector<std::shared_ptr<std::vector<int32_t>>> _cdfs_sizes;
std::vector<std::shared_ptr<std::vector<int32_t>>> _offsets;
std::list<PendingTask> m_pendingEncodingList;
};
class RansEncoderLibMultiThread : public RansEncoderLib {
public:
RansEncoderLibMultiThread();
virtual ~RansEncoderLibMultiThread();
virtual void flush() override;
virtual std::shared_ptr<std::vector<uint8_t>> get_encoded_stream() override;
virtual void reset() override;
void worker();
private:
bool m_finish;
bool m_result_ready;
std::thread m_thread;
std::mutex m_mutex_result;
std::mutex m_mutex_pending;
std::condition_variable m_cv_pending;
std::condition_variable m_cv_result;
std::list<PendingTask> m_pending;
};
class RansDecoderLib {
public:
RansDecoderLib() {}
virtual ~RansDecoderLib() = default;
RansDecoderLib(const RansDecoderLib&) = delete;
RansDecoderLib(RansDecoderLib&&) = delete;
RansDecoderLib& operator=(const RansDecoderLib&) = delete;
RansDecoderLib& operator=(RansDecoderLib&&) = delete;
virtual void set_stream(const std::shared_ptr<std::vector<uint8_t>> encoded);
FORCE_INLINE int8_t decode_one_symbol(const int32_t* cdf, const int32_t cdf_size,
const int32_t offset);
virtual void decode_y(const std::shared_ptr<std::vector<uint8_t>> indexes,
const int cdf_group_index);
virtual void decode_z(const int total_size, const int cdf_group_index, const int start_offset,
const int per_channel_size);
virtual std::shared_ptr<std::vector<int8_t>> get_decoded_tensor();
virtual int add_cdf(const std::shared_ptr<std::vector<std::vector<int32_t>>> cdfs,
const std::shared_ptr<std::vector<int32_t>> cdfs_sizes,
const std::shared_ptr<std::vector<int32_t>> offsets);
virtual void empty_cdf_buffer();
private:
RansState _rans;
uint8_t* _ptr8;
std::shared_ptr<std::vector<uint8_t>> _stream;
std::shared_ptr<std::vector<int8_t>> m_decoded;
std::vector<std::shared_ptr<std::vector<std::vector<int32_t>>>> _cdfs;
std::vector<std::shared_ptr<std::vector<int32_t>>> _cdfs_sizes;
std::vector<std::shared_ptr<std::vector<int32_t>>> _offsets;
};
class RansDecoderLibMultiThread : public RansDecoderLib {
public:
RansDecoderLibMultiThread();
virtual ~RansDecoderLibMultiThread();
virtual void decode_y(const std::shared_ptr<std::vector<uint8_t>> indexes,
const int cdf_group_index) override;
virtual void decode_z(const int total_size, const int cdf_group_index, const int start_offset,
const int per_channel_size) override;
virtual std::shared_ptr<std::vector<int8_t>> get_decoded_tensor() override;
void worker();
private:
bool m_finish;
bool m_result_ready;
std::thread m_thread;
std::mutex m_mutex_result;
std::mutex m_mutex_pending;
std::condition_variable m_cv_pending;
std::condition_variable m_cv_result;
std::list<PendingTask> m_pending;
}; |