MiniCPM-V-4.7#

  • Context Length: 262144

  • Model Name: MiniCPM-V-4.7

  • Languages: en, zh

  • Abilities: chat, vision, hybrid

  • Description: MiniCPM-V 4.7 supports single-image, multi-image, and video understanding with canvas M-RoPE and 4x/16x visual compression. It provides 1B dense and 35B-A3B MoE variants.

Specifications#

Model Spec 1 (pytorch, 1 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 1

  • Quantizations: none

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-1B

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 1 --model-format pytorch --quantization ${quantization}

Model Spec 2 (awq, 1 Billion)#

  • Model Format: awq

  • Model Size (in billions): 1

  • Quantizations: Int4

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-1B-AWQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 1 --model-format awq --quantization ${quantization}

Model Spec 3 (gptq, 1 Billion)#

  • Model Format: gptq

  • Model Size (in billions): 1

  • Quantizations: Int4

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-1B-GPTQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 1 --model-format gptq --quantization ${quantization}

Model Spec 4 (bnb, 1 Billion)#

  • Model Format: bnb

  • Model Size (in billions): 1

  • Quantizations: 4-bit

  • Engines: vLLM, Transformers

  • Model ID: openbmb/MiniCPM-V-4.7-1B-int4

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 1 --model-format bnb --quantization ${quantization}

Model Spec 5 (pytorch, 35 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 35

  • Quantizations: none

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format pytorch --quantization ${quantization}

Model Spec 6 (awq, 35 Billion)#

  • Model Format: awq

  • Model Size (in billions): 35

  • Quantizations: Int4

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B-AWQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format awq --quantization ${quantization}

Model Spec 7 (gptq, 35 Billion)#

  • Model Format: gptq

  • Model Size (in billions): 35

  • Quantizations: Int4

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B-GPTQ

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format gptq --quantization ${quantization}

Model Spec 8 (pytorch, 35 Billion)#

  • Model Format: pytorch

  • Model Size (in billions): 35

  • Quantizations: Int4

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B-int4

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format pytorch --quantization ${quantization}

Model Spec 9 (fp8, 35 Billion)#

  • Model Format: fp8

  • Model Size (in billions): 35

  • Quantizations: FP8

  • Engines: vLLM, Transformers, SGLang

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B-FP8

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format fp8 --quantization ${quantization}

Model Spec 10 (ggufv2, 1 Billion)#

  • Model Format: ggufv2

  • Model Size (in billions): 1

  • Quantizations: Q4_K_M, Q4_K_S, Q5_K_M, Q5_K_S, Q6_K, Q8_0, F16, IQ3_M, IQ3_XS, IQ4_NL, IQ4_XS, Q4_0, Q4_1, Q5_0, Q5_1

  • Engines: llama.cpp

  • Model ID: openbmb/MiniCPM-V-4.7-1B-gguf

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 1 --model-format ggufv2 --quantization ${quantization}

Model Spec 11 (ggufv2, 35 Billion)#

  • Model Format: ggufv2

  • Model Size (in billions): 35

  • Quantizations: Q4_K_M, Q4_K_S, Q5_K_M, Q5_K_S, Q6_K, Q8_0, F16, IQ3_M, IQ3_XS, IQ4_NL, IQ4_XS, Q4_0, Q4_1, Q5_0, Q5_1, mix_q3gud_im

  • Engines: llama.cpp

  • Model ID: openbmb/MiniCPM-V-4.7-35B-A3B-gguf

  • Model Hubs: Hugging Face, ModelScope

Execute the following command to launch the model, remember to replace ${quantization} with your chosen quantization method from the options listed above:

xinference launch --model-engine ${engine} --model-name MiniCPM-V-4.7 --size-in-billions 35 --model-format ggufv2 --quantization ${quantization}