InternLM/InternLM

InternLM

👋 join us on Discord and WeChat

Introduction

InternLM3 has open-sourced an 8-billion parameter instruction model, InternLM3-8B-Instruct, designed for general-purpose usage and advanced reasoning. This model has the following characteristics:

• Enhanced performance at reduced cost: State-of-the-art performance on reasoning and knowledge-intensive tasks surpass models like Llama3.1-8B and Qwen2.5-7B. Remarkably, InternLM3 is trained on only 4 trillion high-quality tokens, saving more than 75% of the training cost compared to other LLMs of similar scale.
• Deep thinking capability: InternLM3 supports both the deep thinking mode for solving complicated reasoning tasks via the long chain-of-thought and the normal response mode for fluent user interactions.

News

[2025.01.15] We release InternLM3-8B-Instruct, See model zoo below for download or model cards for more details.

[2024.08.01] We release InternLM2.5-1.8B, InternLM2.5-1.8B-Chat, InternLM2.5-20B and InternLM2.5-20B-Chat. See model zoo below for download or model cards for more details.

[2024.07.19] We release the InternLM2-Reward series of reward models in 1.8B, 7B and 20B sizes. See model zoo below for download or model cards for more details.

[2024.07.03] We release InternLM2.5-7B, InternLM2.5-7B-Chat and InternLM2.5-7B-Chat-1M. See model zoo below for download or model cards for more details.

[2024.03.26] We release InternLM2 technical report. See arXiv for details.

[2024.01.31] We release InternLM2-1.8B, along with the associated chat model. They provide a cheaper deployment option while maintaining leading performance.

[2024.01.23] We release InternLM2-Math-7B and InternLM2-Math-20B with pretraining and SFT checkpoints. They surpass ChatGPT with small sizes. See InternLM-Math for details and download.

[2024.01.17] We release InternLM2-7B and InternLM2-20B and their corresponding chat models with stronger capabilities in all dimensions. See model zoo below for download or model cards for more details.

[2023.12.13] InternLM-7B-Chat and InternLM-20B-Chat checkpoints are updated. With an improved finetuning strategy, the new chat models can generate higher quality responses with greater stylistic diversity.

[2023.09.20] InternLM-20B is released with base and chat versions.

Model Zoo

InternLM3

InternLM2.5

(click to expand)

Model	Transformers(HF)	ModelScope(HF)	OpenXLab(HF)	OpenXLab(Origin)	Release Date
InternLM2.5-1.8B	🤗internlm2_5-1_8b	internlm2_5-1_8b			2024-08-05
InternLM2.5-1.8B-Chat	🤗internlm2_5-1_8b-chat	internlm2_5-1_8b-chat			2024-08-05
InternLM2.5-7B	🤗internlm2_5-7b	internlm2_5-7b			2024-07-03
InternLM2.5-7B-Chat	🤗internlm2_5-7b-chat	internlm2_5-7b-chat			2024-07-03
InternLM2.5-7B-Chat-1M	🤗internlm2_5-7b-chat-1m	internlm2_5-7b-chat-1m			2024-07-03
InternLM2.5-20B	🤗internlm2_5-20b	internlm2_5-20b			2024-08-05
InternLM2.5-20B-Chat	🤗internlm2_5-20b-chat	internlm2_5-20b-chat			2024-08-05

Notes:

The release of InternLM2.5 series contains 1.8B, 7B, and 20B versions. 7B models are efficient for research and application and 20B models are more powerful and can support more complex scenarios. The relation of these models are shown as follows.

1. InternLM2.5: Foundation models pre-trained on large-scale corpus. InternLM2.5 models are recommended for consideration in most applications.
2. InternLM2.5-Chat: The Chat model that undergoes supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF), based on the InternLM2.5 model. InternLM2.5-Chat is optimized for instruction following, chat experience, and function call, which is recommended for downstream applications.
3. InternLM2.5-Chat-1M: InternLM2.5-Chat-1M supports 1M long-context with compatible performance as InternLM2.5-Chat.

Limitations: Although we have made efforts to ensure the safety of the model during the training process and to encourage the model to generate text that complies with ethical and legal requirements, the model may still produce unexpected outputs due to its size and probabilistic generation paradigm. For example, the generated responses may contain biases, discrimination, or other harmful content. Please do not propagate such content. We are not responsible for any consequences resulting from the dissemination of harmful information.

Supplements: HF refers to the format used by HuggingFace in transformers, whereas Origin denotes the format adopted by the InternLM team in InternEvo.

InternLM2-Reward

(click to expand)

InternLM2-Reward is a series of reward models, trained on 2.4 million preference samples, available in 1.8B, 7B, and 20B sizes. These model were applied to the PPO training process of our chat models. See model cards for more details.

Model	RewardBench Score	Transformers(HF)	ModelScope(HF)	OpenXLab(HF)	Release Date
InternLM2-1.8B-Reward	80.6	🤗internlm2-1_8b-reward	internlm2-1_8b-reward		2024-07-19
InternLM2-7B-Reward	86.6	🤗internlm2-7b-reward	internlm2-7b-reward		2024-07-19
InternLM2-20B-Reward	89.5	🤗internlm2-20b-reward	internlm2-20b-reward		2024-07-19

InternLM2

(click to expand)

Our previous generation models with advanced capabilities in long-context processing, reasoning, and coding. See model cards for more details.

Model	Transformers(HF)	ModelScope(HF)	OpenXLab(HF)	OpenXLab(Origin)	Release Date
InternLM2-1.8B	🤗internlm2-1.8b	internlm2-1.8b			2024-01-31
InternLM2-Chat-1.8B-SFT	🤗internlm2-chat-1.8b-sft	internlm2-chat-1.8b-sft			2024-01-31
InternLM2-Chat-1.8B	🤗internlm2-chat-1.8b	internlm2-chat-1.8b			2024-02-19
InternLM2-Base-7B	🤗internlm2-base-7b	internlm2-base-7b			2024-01-17
InternLM2-7B	🤗internlm2-7b	internlm2-7b			2024-01-17
InternLM2-Chat-7B-SFT	🤗internlm2-chat-7b-sft	internlm2-chat-7b-sft			2024-01-17
InternLM2-Chat-7B	🤗internlm2-chat-7b	internlm2-chat-7b			2024-01-17
InternLM2-Base-20B	🤗internlm2-base-20b	internlm2-base-20b			2024-01-17
InternLM2-20B	🤗internlm2-20b	internlm2-20b			2024-01-17
InternLM2-Chat-20B-SFT	🤗internlm2-chat-20b-sft	internlm2-chat-20b-sft			2024-01-17
InternLM2-Chat-20B	🤗internlm2-chat-20b	internlm2-chat-20b			2024-01-17

Performance

We conducted a comprehensive evaluation of InternLM using the open-source evaluation tool OpenCompass. The evaluation covered five dimensions of capabilities: disciplinary competence, language competence, knowledge competence, inference competence, and comprehension competence. Here are some of the evaluation results, and you can visit the OpenCompass leaderboard for more evaluation results.

• The evaluation results were obtained from OpenCompass (some data marked with *, which means evaluating with Thinking Mode), and evaluation configuration can be found in the configuration files provided by OpenCompass.
• The evaluation data may have numerical differences due to the version iteration of OpenCompass, so please refer to the latest evaluation results of OpenCompass. Limitations: Although we have made efforts to ensure the safety of the model during the training process and to encourage the model to generate text that complies with ethical and legal requirements, the model may still produce unexpected outputs due to its size and probabilistic generation paradigm. For example, the generated responses may contain biases, discrimination, or other harmful content. Please do not propagate such content. We are not responsible for any consequences resulting from the dissemination of harmful information.

Requirements

• Python >= 3.8
• PyTorch >= 1.12.0 (2.0.0 and above are recommended)
• Transformers >= 4.38

Usages

Conversation Mode

Transformers inference

To load the InternLM3 8B Instruct model using Transformers, use the following code:

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_dir = "internlm/internlm3-8b-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_dir, trust_remote_code=True)
# Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
model = AutoModelForCausalLM.from_pretrained(model_dir, trust_remote_code=True, torch_dtype=torch.float16)
# (Optional) If on low resource devices, you can load model in 4-bit or 8-bit to further save GPU memory via bitsandbytes.
  # InternLM3 8B in 4bit will cost nearly 8GB GPU memory.
  # pip install -U bitsandbytes
  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
model = model.eval()
system_prompt = """You are an AI assistant whose name is InternLM (书生·浦语).
- InternLM (书生·浦语) is a conversational language model that is developed by Shanghai AI Laboratory (上海人工智能实验室). It is designed to be helpful, honest, and harmless.
- InternLM (书生·浦语) can understand and communicate fluently in the language chosen by the user such as English and 中文."""
messages = [
    {"role": "system", "content": system_prompt},
    {"role": "user", "content": "Please tell me five scenic spots in Shanghai"},
 ]
tokenized_chat = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
generated_ids = model.generate(tokenized_chat, max_new_tokens=1024, temperature=1, repetition_penalty=1.005, top_k=40, top_p=0.8)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(tokenized_chat, generated_ids)
]
prompt = tokenizer.batch_decode(tokenized_chat)[0]
print(prompt)
response = tokenizer.batch_decode(generated_ids)[0]
print(response)

LMDeploy inference

LMDeploy is a toolkit for compressing, deploying, and serving LLM, developed by the MMRazor and MMDeploy teams.

pip install lmdeploy

You can run batch inference locally with the following python code:

import lmdeploy
model_dir = "internlm/internlm3-8b-instruct"
pipe = lmdeploy.pipeline(model_dir)
response = pipe("Please tell me five scenic spots in Shanghai")
print(response)

Or you can launch an OpenAI compatible server with the following command:

lmdeploy serve api_server internlm/internlm3-8b-instruct --model-name internlm3-8b-instruct --server-port 23333

Then you can send a chat request to the server:

curl http://localhost:23333/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
    "model": "internlm3-8b-instruct",
    "messages": [
    {"role": "user", "content": "Please tell me five scenic spots in Shanghai"}
    ]
    }'

Find more details in the LMDeploy documentation

SGLang inference

Installation

pip3 install "sglang[srt]>=0.4.1.post6" --find-links https://flashinfer.ai/whl/cu124/torch2.4/flashinfer/

OpenAI Compatible Server

python3 -m sglang.launch_server --model internlm/internlm3-8b-instruct --trust-remote-code --chat-template internlm2-chat

OpenAI client

import openai
client = openai.Client(
    base_url="http://127.0.0.1:30000/v1", api_key="EMPTY")

# Chat completion
response = client.chat.completions.create(
    model="default",
    messages=[
        {"role": "system", "content": "You are a helpful AI assistant"},
        {"role": "user", "content": "List 3 countries and their capitals."},
    ],
    temperature=0,
    max_tokens=64,
)
print(response)

Ollama inference

install ollama and pull the model

# install ollama
curl -fsSL https://ollama.com/install.sh | sh
# pull the model
ollama pull internlm/internlm3-8b-instruct
# install ollama-python
pip install ollama

inference code:

import ollama

system_prompt = """You are an AI assistant whose name is InternLM (书生·浦语).
- InternLM (书生·浦语) is a conversational language model that is developed by Shanghai AI Laboratory (上海人工智能实验室). It is designed to be helpful, honest, and harmless.
- InternLM (书生·浦语) can understand and communicate fluently in the language chosen by the user such as English and 中文."""

messages = [
    {
        "role": "system",
        "content": system_prompt,
    },
    {
        "role": "user",
        "content": "Please tell me five scenic spots in Shanghai"
    },
]

stream = ollama.chat(
    model='internlm/internlm3-8b-instruct',
    messages=messages,
    stream=True,
)

for chunk in stream:
  print(chunk['message']['content'], end='', flush=True)

vLLM inference

refer to installation to install the latest code of vllm

pip install vllm --pre --extra-index-url https://wheels.vllm.ai/nightly

inference code:

from vllm import LLM, SamplingParams
llm = LLM(model="internlm/internlm3-8b-instruct")
sampling_params = SamplingParams(temperature=1, repetition_penalty=1.005, top_k=40, top_p=0.8)
system_prompt = """You are an AI assistant whose name is InternLM (书生·浦语).
- InternLM (书生·浦语) is a conversational language model that is developed by Shanghai AI Laboratory (上海人工智能实验室). It is designed to be helpful, honest, and harmless.
- InternLM (书生·浦语) can understand and communicate fluently in the language chosen by the user such as English and 中文."""
prompts = [
    {
        "role": "system",
        "content": system_prompt,
    },
    {
        "role": "user",
        "content": "Please tell me five scenic spots in Shanghai"
    },
]
outputs = llm.chat(prompts,
                   sampling_params=sampling_params,
                   use_tqdm=False)
print(outputs)

Thinking Mode

Thinking Demo

Thinking system prompt

thinking_system_prompt = """You are an expert mathematician with extensive experience in mathematical competitions. You approach problems through systematic thinking and rigorous reasoning. When solving problems, follow these thought processes:
## Deep Understanding
Take time to fully comprehend the problem before attempting a solution. Consider:
- What is the real question being asked?
- What are the given conditions and what do they tell us?
- Are there any special restrictions or assumptions?
- Which information is crucial and which is supplementary?
## Multi-angle Analysis
Before solving, conduct thorough analysis:
- What mathematical concepts and properties are involved?
- Can you recall similar classic problems or solution methods?
- Would diagrams or tables help visualize the problem?
- Are there special cases that need separate consideration?
## Systematic Thinking
Plan your solution path:
- Propose multiple possible approaches
- Analyze the feasibility and merits of each method
- Choose the most appropriate method and explain why
- Break complex problems into smaller, manageable steps
## Rigorous Proof
During the solution process:
- Provide solid justification for each step
- Include detailed proofs for key conclusions
- Pay attention to logical connections
- Be vigilant about potential oversights
## Repeated Verification
After completing your solution:
- Verify your results satisfy all conditions
- Check for overlooked special cases
- Consider if the solution can be optimized or simplified
- Review your reasoning process
Remember:
1. Take time to think thoroughly rather than rushing to an answer
2. Rigorously prove each key conclusion
3. Keep an open mind and try different approaches
4. Summarize valuable problem-solving methods
5. Maintain healthy skepticism and verify multiple times
Your response should reflect deep mathematical understanding and precise logical thinking, making your solution path and reasoning clear to others.
When you're ready, present your complete solution with:
- Clear problem understanding
- Detailed solution process
- Key insights
- Thorough verification
Focus on clear, logical progression of ideas and thorough explanation of your mathematical reasoning. Provide answers in the same language as the user asking the question, repeat the final answer using a '\\boxed{}' without any units, you have [[8192]] tokens to complete the answer.
"""

Transformers inference

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_dir = "internlm/internlm3-8b-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_dir, trust_remote_code=True)
# Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
model = AutoModelForCausalLM.from_pretrained(model_dir, trust_remote_code=True, torch_dtype=torch.float16)
# (Optional) If on low resource devices, you can load model in 4-bit or 8-bit to further save GPU memory via bitsandbytes.
  # InternLM3 8B in 4bit will cost nearly 8GB GPU memory.
  # pip install -U bitsandbytes
  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
model = model.eval()
messages = [
    {"role": "system", "content": thinking_system_prompt},
    {"role": "user", "content": "Given the function\(f(x)=\mathrm{e}^{x}-ax - a^{3}\),\n(1) When \(a = 1\), find the equation of the tangent line to the curve \(y = f(x)\) at the point \((1,f(1))\).\n(2) If \(f(x)\) has a local minimum and the minimum value is less than \(0\), determine the range of values for \(a\)."},
 ]
tokenized_chat = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
generated_ids = model.generate(tokenized_chat, max_new_tokens=8192)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(tokenized_chat, generated_ids)
]
prompt = tokenizer.batch_decode(tokenized_chat)[0]
print(prompt)
response = tokenizer.batch_decode(generated_ids)[0]
print(response)

LMDeploy inference

LMDeploy is a toolkit for compressing, deploying, and serving LLM.

pip install lmdeploy

You can run batch inference locally with the following python code:

from lmdeploy import pipeline, GenerationConfig, ChatTemplateConfig
model_dir = "internlm/internlm3-8b-instruct"
chat_template_config = ChatTemplateConfig(model_name='internlm3')
pipe = pipeline(model_dir, chat_template_config=chat_template_config)
messages = [
        {"role": "system", "content": thinking_system_prompt},
        {"role": "user", "content": "Given the function\(f(x)=\mathrm{e}^{x}-ax - a^{3}\),\n(1) When \(a = 1\), find the equation of the tangent line to the curve \(y = f(x)\) at the point \((1,f(1))\).\n(2) If \(f(x)\) has a local minimum and the minimum value is less than \(0\), determine the range of values for \(a\)."},
]
response = pipe(messages, gen_config=GenerationConfig(max_new_tokens=2048))
print(response)

SGLang inference

Installation

pip3 install "sglang[srt]>=0.4.1.post6" --find-links https://flashinfer.ai/whl/cu124/torch2.4/flashinfer/

For offline engine api usage, please refer to Offline Engine API

Ollama inference

install ollama and pull the model

# install ollama
curl -fsSL https://ollama.com/install.sh | sh
# pull the model
ollama pull internlm/internlm3-8b-instruct
# install ollama-python
pip install ollama

inference code:

import ollama

messages = [
    {
        "role": "system",
        "content": thinking_system_prompt,
    },
    {
        "role": "user",
        "content": "已知函数\(f(x)=\mathrm{e}^{x}-ax - a^{3}\)。\n（1）当\(a = 1\)时，求曲线\(y = f(x)\)在点\((1,f(1))\)处的切线方程；\n（2）若\(f(x)\)有极小值，且极小值小于\(0\)，求\(a\)的取值范围。"
    },
]

stream = ollama.chat(
    model='internlm/internlm3-8b-instruct',
    messages=messages,
    stream=True,
    options=dict(num_ctx=8192, num_predict=2048)
)

for chunk in stream:
  print(chunk['message']['content'], end='', flush=True)

vLLM inference

refer to installation to install the latest code of vllm

pip install vllm --pre --extra-index-url https://wheels.vllm.ai/nightly

inference code

from vllm import LLM, SamplingParams
llm = LLM(model="internlm/internlm3-8b-instruct")
sampling_params = SamplingParams(temperature=1, repetition_penalty=1.005, top_k=40, top_p=0.8, max_tokens=8192)
prompts = [
    {
        "role": "system",
        "content": thinking_system_prompt,
    },
    {
        "role": "user",
        "content": "Given the function\(f(x)=\mathrm{e}^{x}-ax - a^{3}\),\n(1) When \(a = 1\), find the equation of the tangent line to the curve \(y = f(x)\) at the point \((1,f(1))\).\n(2) If \(f(x)\) has a local minimum and the minimum value is less than \(0\), determine the range of values for \(a\)."
    },
]
outputs = llm.chat(prompts,
                   sampling_params=sampling_params,
                   use_tqdm=False)
print(outputs)

License

Code and model weights are licensed under Apache-2.0.

Citation

@misc{cai2024internlm2,
      title={InternLM2 Technical Report},
      author={Zheng Cai and Maosong Cao and Haojiong Chen and Kai Chen and Keyu Chen and Xin Chen and Xun Chen and Zehui Chen and Zhi Chen and Pei Chu and Xiaoyi Dong and Haodong Duan and Qi Fan and Zhaoye Fei and Yang Gao and Jiaye Ge and Chenya Gu and Yuzhe Gu and Tao Gui and Aijia Guo and Qipeng Guo and Conghui He and Yingfan Hu and Ting Huang and Tao Jiang and Penglong Jiao and Zhenjiang Jin and Zhikai Lei and Jiaxing Li and Jingwen Li and Linyang Li and Shuaibin Li and Wei Li and Yining Li and Hongwei Liu and Jiangning Liu and Jiawei Hong and Kaiwen Liu and Kuikun Liu and Xiaoran Liu and Chengqi Lv and Haijun Lv and Kai Lv and Li Ma and Runyuan Ma and Zerun Ma and Wenchang Ning and Linke Ouyang and Jiantao Qiu and Yuan Qu and Fukai Shang and Yunfan Shao and Demin Song and Zifan Song and Zhihao Sui and Peng Sun and Yu Sun and Huanze Tang and Bin Wang and Guoteng Wang and Jiaqi Wang and Jiayu Wang and Rui Wang and Yudong Wang and Ziyi Wang and Xingjian Wei and Qizhen Weng and Fan Wu and Yingtong Xiong and Chao Xu and Ruiliang Xu and Hang Yan and Yirong Yan and Xiaogui Yang and Haochen Ye and Huaiyuan Ying and Jia Yu and Jing Yu and Yuhang Zang and Chuyu Zhang and Li Zhang and Pan Zhang and Peng Zhang and Ruijie Zhang and Shuo Zhang and Songyang Zhang and Wenjian Zhang and Wenwei Zhang and Xingcheng Zhang and Xinyue Zhang and Hui Zhao and Qian Zhao and Xiaomeng Zhao and Fengzhe Zhou and Zaida Zhou and Jingming Zhuo and Yicheng Zou and Xipeng Qiu and Yu Qiao and Dahua Lin},
      year={2024},
      eprint={2403.17297},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}