From 789b84bb38a7cb9eda1e79eee51c4f58f9ae6c47 Mon Sep 17 00:00:00 2001 From: ModelHub XC Date: Thu, 30 Jul 2026 16:41:17 +0800 Subject: [PATCH] =?UTF-8?q?=E5=88=9D=E5=A7=8B=E5=8C=96=E9=A1=B9=E7=9B=AE?= =?UTF-8?q?=EF=BC=8C=E7=94=B1ModelHub=20XC=E7=A4=BE=E5=8C=BA=E6=8F=90?= =?UTF-8?q?=E4=BE=9B=E6=A8=A1=E5=9E=8B?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Model: mykor/Midm-2.0-Base-Instruct-gguf Source: Original Platform --- .gitattributes | 36 ++ LICENSE.txt | 21 ++ Midm-2.0-Base-Instruct-BF16.gguf | 3 + Midm-2.0-Base-Instruct-F32.gguf | 3 + Midm-2.0-Base-Instruct-IQ4_NL.gguf | 3 + Midm-2.0-Base-Instruct-IQ4_XS.gguf | 3 + Midm-2.0-Base-Instruct-Q2_K.gguf | 3 + Midm-2.0-Base-Instruct-Q3_K_L.gguf | 3 + Midm-2.0-Base-Instruct-Q3_K_M.gguf | 3 + Midm-2.0-Base-Instruct-Q3_K_S.gguf | 3 + Midm-2.0-Base-Instruct-Q4_0.gguf | 3 + Midm-2.0-Base-Instruct-Q4_K_M.gguf | 3 + Midm-2.0-Base-Instruct-Q4_K_S.gguf | 3 + Midm-2.0-Base-Instruct-Q5_0.gguf | 3 + Midm-2.0-Base-Instruct-Q5_K_M.gguf | 3 + Midm-2.0-Base-Instruct-Q5_K_S.gguf | 3 + Midm-2.0-Base-Instruct-Q6_K.gguf | 3 + Midm-2.0-Base-Instruct-Q8_0.gguf | 3 + README.md | 575 +++++++++++++++++++++++++++++ 19 files changed, 680 insertions(+) create mode 100644 .gitattributes create mode 100644 LICENSE.txt create mode 100644 Midm-2.0-Base-Instruct-BF16.gguf create mode 100644 Midm-2.0-Base-Instruct-F32.gguf create mode 100644 Midm-2.0-Base-Instruct-IQ4_NL.gguf create mode 100644 Midm-2.0-Base-Instruct-IQ4_XS.gguf create mode 100644 Midm-2.0-Base-Instruct-Q2_K.gguf create mode 100644 Midm-2.0-Base-Instruct-Q3_K_L.gguf create mode 100644 Midm-2.0-Base-Instruct-Q3_K_M.gguf create mode 100644 Midm-2.0-Base-Instruct-Q3_K_S.gguf create mode 100644 Midm-2.0-Base-Instruct-Q4_0.gguf create mode 100644 Midm-2.0-Base-Instruct-Q4_K_M.gguf create mode 100644 Midm-2.0-Base-Instruct-Q4_K_S.gguf create mode 100644 Midm-2.0-Base-Instruct-Q5_0.gguf create mode 100644 Midm-2.0-Base-Instruct-Q5_K_M.gguf create mode 100644 Midm-2.0-Base-Instruct-Q5_K_S.gguf create mode 100644 Midm-2.0-Base-Instruct-Q6_K.gguf create mode 100644 Midm-2.0-Base-Instruct-Q8_0.gguf create mode 100644 README.md diff --git a/.gitattributes b/.gitattributes new file mode 100644 index 0000000..f15b49c --- /dev/null +++ b/.gitattributes @@ -0,0 +1,36 @@ +*.7z filter=lfs diff=lfs merge=lfs -text +*.arrow filter=lfs diff=lfs merge=lfs -text +*.bin filter=lfs diff=lfs merge=lfs -text +*.bz2 filter=lfs diff=lfs merge=lfs -text +*.ckpt filter=lfs diff=lfs merge=lfs -text +*.ftz filter=lfs diff=lfs merge=lfs -text +*.gz filter=lfs diff=lfs merge=lfs -text +*.h5 filter=lfs diff=lfs merge=lfs -text +*.joblib filter=lfs diff=lfs merge=lfs -text +*.lfs.* filter=lfs diff=lfs merge=lfs -text +*.mlmodel filter=lfs diff=lfs merge=lfs -text +*.model filter=lfs diff=lfs merge=lfs -text +*.msgpack filter=lfs diff=lfs merge=lfs -text +*.npy filter=lfs diff=lfs merge=lfs -text +*.npz filter=lfs diff=lfs merge=lfs -text +*.onnx filter=lfs diff=lfs merge=lfs -text +*.ot filter=lfs diff=lfs merge=lfs -text +*.parquet filter=lfs diff=lfs merge=lfs -text +*.pb filter=lfs diff=lfs merge=lfs -text +*.pickle filter=lfs diff=lfs merge=lfs -text +*.pkl filter=lfs diff=lfs merge=lfs -text +*.pt filter=lfs diff=lfs merge=lfs -text +*.pth filter=lfs diff=lfs merge=lfs -text +*.rar filter=lfs diff=lfs merge=lfs -text +*.safetensors filter=lfs diff=lfs merge=lfs -text +saved_model/**/* filter=lfs diff=lfs merge=lfs -text +*.tar.* filter=lfs diff=lfs merge=lfs -text +*.tar filter=lfs diff=lfs merge=lfs -text +*.tflite filter=lfs diff=lfs merge=lfs -text +*.tgz filter=lfs diff=lfs merge=lfs -text +*.wasm filter=lfs diff=lfs merge=lfs -text +*.xz filter=lfs diff=lfs merge=lfs -text +*.zip filter=lfs diff=lfs merge=lfs -text +*.zst filter=lfs diff=lfs merge=lfs -text +*tfevents* filter=lfs diff=lfs merge=lfs -text +*.gguf filter=lfs diff=lfs merge=lfs -text diff --git a/LICENSE.txt b/LICENSE.txt new file mode 100644 index 0000000..df5e8f4 --- /dev/null +++ b/LICENSE.txt @@ -0,0 +1,21 @@ +MIT License + +Copyright (c) 2025 KT Corporation + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. diff --git a/Midm-2.0-Base-Instruct-BF16.gguf b/Midm-2.0-Base-Instruct-BF16.gguf new file mode 100644 index 0000000..ea2181e --- /dev/null +++ b/Midm-2.0-Base-Instruct-BF16.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b74f96d96cbfaa878b89310689e7b9a33db835c5bfff1309cdf3d4c9b2fc697 +size 23097821056 diff --git a/Midm-2.0-Base-Instruct-F32.gguf b/Midm-2.0-Base-Instruct-F32.gguf new file mode 100644 index 0000000..2ad8e23 --- /dev/null +++ b/Midm-2.0-Base-Instruct-F32.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b67cc294e38748026582a29915515d120404850d556218d786b901463e7d5d38 +size 46188382080 diff --git a/Midm-2.0-Base-Instruct-IQ4_NL.gguf b/Midm-2.0-Base-Instruct-IQ4_NL.gguf new file mode 100644 index 0000000..c64a055 --- /dev/null +++ b/Midm-2.0-Base-Instruct-IQ4_NL.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3c94b09fa90cf6643000238399a3e80c894d17d90685bfa72bc41406bd2d29b8 +size 6709427840 diff --git a/Midm-2.0-Base-Instruct-IQ4_XS.gguf b/Midm-2.0-Base-Instruct-IQ4_XS.gguf new file mode 100644 index 0000000..0f5865e --- /dev/null +++ b/Midm-2.0-Base-Instruct-IQ4_XS.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10b011395362ad4095a77387b656e8aba5d02f69ab5568da0dd14576971c860a +size 6382756480 diff --git a/Midm-2.0-Base-Instruct-Q2_K.gguf b/Midm-2.0-Base-Instruct-Q2_K.gguf new file mode 100644 index 0000000..c0bc58d --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q2_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9677673f86b541ccb2ce2f38f3dc7269dab46164a9553c03da6fac117b688e55 +size 4475661440 diff --git a/Midm-2.0-Base-Instruct-Q3_K_L.gguf b/Midm-2.0-Base-Instruct-Q3_K_L.gguf new file mode 100644 index 0000000..849fc14 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q3_K_L.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a399c916435d22dfbaf1b1ac1a5690af73f4a544fc1a3bd1031011ceea4536cd +size 6164522880 diff --git a/Midm-2.0-Base-Instruct-Q3_K_M.gguf b/Midm-2.0-Base-Instruct-Q3_K_M.gguf new file mode 100644 index 0000000..89423fb --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q3_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8dac1dcde0fa6dcdf7d17f243b7bffc44417b6a7855f298670e45fb5a08988ce +size 5709440896 diff --git a/Midm-2.0-Base-Instruct-Q3_K_S.gguf b/Midm-2.0-Base-Instruct-Q3_K_S.gguf new file mode 100644 index 0000000..79d1e59 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q3_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:44e2f9dbed57861921a8d29d04348b974014cb1ea8d37680220df6e2fa96b1dc +size 5178337152 diff --git a/Midm-2.0-Base-Instruct-Q4_0.gguf b/Midm-2.0-Base-Instruct-Q4_0.gguf new file mode 100644 index 0000000..2098725 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q4_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:70dd57cf11dc2d222412f8d9996ef05236881d1b16ee067ba38004a69319ddf3 +size 6640221824 diff --git a/Midm-2.0-Base-Instruct-Q4_K_M.gguf b/Midm-2.0-Base-Instruct-Q4_K_M.gguf new file mode 100644 index 0000000..af618a7 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q4_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34f5e8107a1469ac7cc6358b68e84e81a32f77d4cfcfba8be7f83dc4385f1d8b +size 7029505664 diff --git a/Midm-2.0-Base-Instruct-Q4_K_S.gguf b/Midm-2.0-Base-Instruct-Q4_K_S.gguf new file mode 100644 index 0000000..3005d7e --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q4_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:472cb492e4c9bb54031528c912572d9ae3dc76f6254073f239e17839bf4ba983 +size 6686359168 diff --git a/Midm-2.0-Base-Instruct-Q5_0.gguf b/Midm-2.0-Base-Instruct-Q5_0.gguf new file mode 100644 index 0000000..132e6ae --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q5_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:02f18417cc91d7ced0c307b24f3fc990f22cef2d520469538b24b89fae5d8a8b +size 8016113280 diff --git a/Midm-2.0-Base-Instruct-Q5_K_M.gguf b/Midm-2.0-Base-Instruct-Q5_K_M.gguf new file mode 100644 index 0000000..7ac4529 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q5_K_M.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c470240fb3578906d111ce8ed60f4b7d41b2771612214fc2cc228a6eec38e10c +size 8216653440 diff --git a/Midm-2.0-Base-Instruct-Q5_K_S.gguf b/Midm-2.0-Base-Instruct-Q5_K_S.gguf new file mode 100644 index 0000000..a398dac --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q5_K_S.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38463bc5ac25d6c8dcc5a2ea4b52d69c4aea508cd4b9ce98ba51b74e30f74a5f +size 8016113280 diff --git a/Midm-2.0-Base-Instruct-Q6_K.gguf b/Midm-2.0-Base-Instruct-Q6_K.gguf new file mode 100644 index 0000000..87c6c65 --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q6_K.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:911ad76f01a89a6659aeffd2290a711f4395019facd96abc4e9085f42316f34a +size 9477997952 diff --git a/Midm-2.0-Base-Instruct-Q8_0.gguf b/Midm-2.0-Base-Instruct-Q8_0.gguf new file mode 100644 index 0000000..090699b --- /dev/null +++ b/Midm-2.0-Base-Instruct-Q8_0.gguf @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29c5ecc48d805ea6f4f2fe79d6a7691332cee1c783a2d30fd88d5f3ef6687ae5 +size 12274120576 diff --git a/README.md b/README.md new file mode 100644 index 0000000..48eb6c3 --- /dev/null +++ b/README.md @@ -0,0 +1,575 @@ +--- +license: mit +language: +- en +- ko +tags: +- KT +- K-intelligence +- Mi:dm +inference: true +pipeline_tag: text-generation +base_model: +- K-intelligence/Midm-2.0-Base-Instruct +--- + + +

+ + + +

+ +

+๐Ÿค— Mi:dm 2.0 Models | +๐Ÿ“œ Mi:dm 2.0 Technical Report | +๐Ÿ“• Mi:dm 2.0 Technical Blog* +

+ +

*To be released soon

+ +
+ +# News ๐Ÿ“ข + +- ๐Ÿ”œ _(Coming Soon!) GGUF format model files will be available soon for easier local deployment._ +- โšก๏ธ`2025/07/04`: Released Mi:dm 2.0 Model collection on Hugging Face๐Ÿค—. +
+
+# Table of Contents + +- ___Overview___ + - [Mi:dm 2.0](#midm-20) + - [Quickstart](#quickstart) + - [Evaluation](#evaluation) +- ___Usage___ + - [Run on Friendli.AI](#run-on-friendliai) + - [Run on Your Local Machine](#run-on-your-local-machine) + - [Deployment](#deployment) + - [Tutorials](#tutorials) +- ___More Information___ + - [Limitation](#limitation) + - [License](#license) + - [Contact](#contact) + +
+
+ +# Overview + +### Mi:dm 2.0 + +**Mi:dm 2.0** is a __"Korea-centric AI"__ model developed using KT's proprietary technology. The term __"Korea-centric AI"__ refers to a model that deeply internalizes the unique values, cognitive frameworks, and commonsense reasoning inherent to Korean society. It goes beyond simply processing or generating Korean textโ€”it reflects a deeper understanding of the socio-cultural norms and values that define Korean society. + +Mi:dm 2.0 is released in two versions: + +- **Mi:dm 2.0 Base** + An 11.5B parameter dense model designed to balance model size and performance. + It extends an 8B-scale model by applying the Depth-up Scaling (DuS) method, making it suitable for real-world applications that require both performance and versatility. + +- **Mi:dm 2.0 Mini** + A lightweight 2.3B parameter dense model optimized for on-device environments and systems with limited GPU resources. + It was derived from the Base model through pruning and distillation to enable compact deployment. + +> [!Note] +> Neither the pre-training nor the post-training data includes KT users' data. + +
+ +### Quickstart + +Here is the code snippet to run conversational inference with the model: + +```python +import torch +from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig + +model_name = "K-intelligence/Midm-2.0-Base-Instruct" + +model = AutoModelForCausalLM.from_pretrained( + model_name, + torch_dtype=torch.bfloat16, + trust_remote_code=True, + device_map="auto" +) +tokenizer = AutoTokenizer.from_pretrained(model_name) +generation_config = GenerationConfig.from_pretrained(model_name) + +prompt = "KT์— ๋Œ€ํ•ด ์†Œ๊ฐœํ•ด์ค˜" + +# message for inference +messages = [ + {"role": "system", + "content": "Mi:dm(๋ฏฟ:์Œ)์€ KT์—์„œ ๊ฐœ๋ฐœํ•œ AI ๊ธฐ๋ฐ˜ ์–ด์‹œ์Šคํ„ดํŠธ์ด๋‹ค."}, + {"role": "user", "content": prompt} +] + +input_ids = tokenizer.apply_chat_template( + messages, + tokenize=True, + add_generation_prompt=True, + return_tensors="pt" +) + +output = model.generate( + input_ids.to("cuda"), + generation_config=generation_config, + eos_token_id=tokenizer.eos_token_id, + max_new_tokens=128, + do_sample=False, +) +print(tokenizer.decode(output[0])) +``` + +> [!NOTE] +> The `transformers` library should be version `4.45.0` or higher. + +
+ +# Evaluation + + +#### Korean + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelSociety & CultureGeneral KnowledgeInstruction Following
K-Refer*K-Refer-Hard*Ko-Sovereign*HAERAEAvg.KMMLUKo-Sovereign*Avg.Ko-IFEvalKo-MTBenchAvg.
Qwen3-4B53.642.935.850.645.750.642.546.575.963.069.4
Exaone-3.5-2.4B-inst64.067.144.461.359.243.542.443.065.474.068.9
Mi:dm 2.0-Mini-inst66.461.436.770.858.845.142.443.873.374.073.6
Qwen3-14B72.465.749.868.464.155.454.755.183.67177.3
Llama-3.1-8B-inst43.236.433.849.540.733.036.734.860.15758.5
Exaone-3.5-7.8B-inst71.669.346.972.965.252.645.649.169.179.674.4
Mi:dm 2.0-Base-inst89.686.456.381.578.457.358.057.78289.785.9
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelComprehensionReasoning
K-Prag*K-Refer-Hard*Ko-BestKo-Sovereign*Avg.Ko-WinograndeKo-BestLogicKorHRM8KAvg.
Qwen3-4B73.956.791.543.566.667.569.25.656.743.8
Exaone-3.5-2.4B-inst68.758.587.238.062.560.364.17.438.536.7
Mi:dm 2.0-Mini-inst69.555.480.542.561.961.764.57.739.937.4
Qwen3-14B86.774.093.952.076.877.275.46.464.548.8
Llama-3.1-8B-inst59.948.677.431.551.540.126.02.430.919.8
Exaone-3.5-7.8B-inst73.561.992.044.067.264.660.38.649.739.5
Mi:dm 2.0-Base-inst86.570.895.253.076.175.173.08.652.944.8
+ +`*` indicates KT proprietary evaluation resources. + +
+ + +#### English + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ModelInstructionReasoningMathCodingGeneral Knowledge
IFEvalBBHGPQAMuSRAvg.GSM8KMBPP+MMLU-proMMLUAvg.
Qwen3-4B79.779.039.858.559.190.462.4-73.373.3
Exaone-3.5-2.4B-inst81.146.428.149.741.482.559.8-59.559.5
Mi:dm 2.0-Mini-inst73.644.526.651.740.983.160.9-56.556.5
 
Qwen3-14B83.983.449.857.763.688.073.470.582.776.6
Llama-3.1-8B-inst79.960.321.650.344.181.281.847.670.759.2
Exaone-3.5-7.8B-inst83.650.133.151.244.881.179.440.769.054.8
Mi:dm 2.0-Base-inst84.077.733.551.954.491.677.553.373.763.5
+ + +
+ +# Usage + +### Run on Friendli.AI +You can try our model immediately via `Friendli.AI`. Simply click `Deploy` and then `Friendli Endpoints`. + +> [!Note] +> Please note that a login to `Friendli.AI` is required after your fifth chat interaction. + +

+ Left Image + Right Image +

+ + +### Run on Your Local Machine +We provide a detailed description about running Mi:dm 2.0 on your local machine using llama.cpp, LM Studio, and Ollama. Please check our [github](https://github.com/K-intelligence-Midm/Midm-2.0) for more information + + +### Deployment + +To serve Mi:dm 2.0 using [vLLM](https://github.com/vllm-project/vllm)(`>=0.8.0`) with an OpenAI-compatible API: +```bash +vllm serve K-intelligence/Midm-2.0-Base-Instruct +``` + + +### Tutorials +To help our end-users easily use Mi:dm 2.0, we have provided comprehensive tutorials on [github](https://github.com/K-intelligence-Midm/Midm-2.0). +
+ +
+
+ +# More Information + +### Limitation +* The training data for both Mi:dm 2.0 models consists primarily of English and Korean. Understanding and generation in other languages are not guaranteed. + +* The model is not guaranteed to provide reliable advice in fields that require professional expertise, such as law, medicine, or finance. + +* Researchers have made efforts to exclude unethical content from the training data โ€” such as profanity, slurs, bias, and discriminatory language. However, despite these efforts, the model may still produce inappropriate expressions or factual inaccuracies. + + +### License + +Mi:dm 2.0 is licensed under the [MIT License](./LICENSE). + + +### Contact +Mi:dm 2.0 Technical Inquiries: midm-llm@kt.com + +
\ No newline at end of file