Upload folder using huggingface_hub
Browse files- LICENSE +96 -0
- README.md +177 -0
- Solar-Open-100B.png +0 -0
- generation_config.json +4 -4
- tokenizer_config.json +3 -5
LICENSE
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Solar-Apache License, Version 2.0
|
| 2 |
+
|
| 3 |
+
Preamble
|
| 4 |
+
|
| 5 |
+
The 'Solar-Apache License Version 2.0' (hereinafter referred to as "this License") was established by Upstage Inc., incorporated under the laws of the Republic of Korea, to encourage the development of open-source software using Solar AI models, and is not affiliated with the Apache Software Foundation.
|
| 6 |
+
|
| 7 |
+
This License basically adopts all provisions of the Apache License, Version 2.0 (hereinafter referred to as "Apache License 2.0"), including the principle of allowing commercial use, but prescribes minimum strategic conditions for the global expansion of AI technology and the sustainable development of the AI ecosystem.
|
| 8 |
+
|
| 9 |
+
The key additional condition, as specified in Section 4(e), is that if you distribute a "Derivative AI Model" based on the "Work", you must specify the 'Solar' brand. This applies as an exception to Section 6 (Trademarks) of the Apache License 2.0.
|
| 10 |
+
|
| 11 |
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
| 12 |
+
|
| 13 |
+
1. Definitions.
|
| 14 |
+
|
| 15 |
+
"License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document.
|
| 16 |
+
|
| 17 |
+
"Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License.
|
| 18 |
+
|
| 19 |
+
"Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity.
|
| 20 |
+
|
| 21 |
+
"You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License.
|
| 22 |
+
|
| 23 |
+
"Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files.
|
| 24 |
+
|
| 25 |
+
"Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types.
|
| 26 |
+
|
| 27 |
+
"Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below).
|
| 28 |
+
|
| 29 |
+
"Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof.
|
| 30 |
+
|
| 31 |
+
"Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution."
|
| 32 |
+
|
| 33 |
+
"Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work.
|
| 34 |
+
|
| 35 |
+
2. Grant of Copyright License.
|
| 36 |
+
Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form.
|
| 37 |
+
|
| 38 |
+
3. Grant of Patent License.
|
| 39 |
+
Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed.
|
| 40 |
+
|
| 41 |
+
4. Redistribution.
|
| 42 |
+
You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions:
|
| 43 |
+
|
| 44 |
+
(a) You must give any other recipients of the Work or Derivative Works a copy of this License; and
|
| 45 |
+
(b) You must cause any modified files to carry prominent notices stating that You changed the files; and
|
| 46 |
+
(c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and
|
| 47 |
+
(d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License.
|
| 48 |
+
(e) If You distribute or make available a Derivative Work that is an artificial intelligence model created, trained, fine-tuned, or otherwise improved using the Work (the "Derivative AI Model"), You must adhere to the following conditions:
|
| 49 |
+
(i) The name of such Derivative AI Model must begin with "Solar" (e.g., "Solar-MyModel-v1"); and
|
| 50 |
+
(ii) You must prominently display the phrase "Built with Solar" in any related websites, user interfaces, or documentation associated with the Derivative AI Model; and
|
| 51 |
+
(iii) You must provide a copy of this License, including the original copyright notice and NOTICE file as included in the distribution of the Works, alongside the Derivative AI Model.
|
| 52 |
+
|
| 53 |
+
You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License.
|
| 54 |
+
|
| 55 |
+
5. Submission of Contributions.
|
| 56 |
+
Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions.
|
| 57 |
+
|
| 58 |
+
6. Trademarks.
|
| 59 |
+
This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work, reproducing the content of the NOTICE file, and as explicitly required for attribution in Section 4(e) of this License.
|
| 60 |
+
|
| 61 |
+
7. Disclaimer of Warranty.
|
| 62 |
+
Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License.
|
| 63 |
+
|
| 64 |
+
8. Limitation of Liability.
|
| 65 |
+
In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages.
|
| 66 |
+
|
| 67 |
+
9. Accepting Warranty or Additional Liability.
|
| 68 |
+
While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability.
|
| 69 |
+
|
| 70 |
+
END OF TERMS AND CONDITIONS
|
| 71 |
+
|
| 72 |
+
APPENDIX: How to apply the Solar-Apache License to your work.
|
| 73 |
+
|
| 74 |
+
To apply the Solar-Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives.
|
| 75 |
+
|
| 76 |
+
Copyright [yyyy] [Upstage AI (or other copyright owner)]
|
| 77 |
+
|
| 78 |
+
Licensed under the Solar-Apache License, Version 2.0 (the "License");
|
| 79 |
+
you may not use this file except in compliance with the License.
|
| 80 |
+
You may obtain a copy of the License at
|
| 81 |
+
|
| 82 |
+
https://huggingface.co/Upstage/Solar-Open-100B/blob/main/LICENSE
|
| 83 |
+
|
| 84 |
+
Unless required by applicable law or agreed to in writing, software
|
| 85 |
+
distributed under the License is distributed on an "AS IS" BASIS,
|
| 86 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 87 |
+
See the License for the specific language governing permissions and
|
| 88 |
+
limitations under the License.
|
| 89 |
+
|
| 90 |
+
This work is based on or derived from materials licensed under the
|
| 91 |
+
Solar-Apache License, Version 2.0.
|
| 92 |
+
If you distribute or make available a Derivative AI Model (as defined
|
| 93 |
+
in Section 4(e) of the License), your model name must begin with "Solar"
|
| 94 |
+
and you must prominently display "Built with Solar" in associated
|
| 95 |
+
documentation or interfaces.
|
| 96 |
+
|
README.md
ADDED
|
@@ -0,0 +1,177 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
- ko
|
| 5 |
+
license: other
|
| 6 |
+
license_name: solar-apache-2.0
|
| 7 |
+
tags:
|
| 8 |
+
- upstage
|
| 9 |
+
- solar
|
| 10 |
+
- moe
|
| 11 |
+
- 100b
|
| 12 |
+
- llm
|
| 13 |
+
base_model: upstage/Solar-Open-100B
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<p align="center">
|
| 17 |
+
<img src="./Solar-Open-100B.png" alt="Solar Open Model" width="100%">
|
| 18 |
+
</p>
|
| 19 |
+
|
| 20 |
+
# **Solar Open**
|
| 21 |
+
|
| 22 |
+
**Solar Open** is Upstage's flagship **102B-parameter** large language model, trained **entirely from scratch** and released under the **Solar-Apache License 2.0** (see [LICENSE](./LICENSE)). As a **Mixture-of-Experts (MoE)** architecture, it delivers enterprise-grade performance in reasoning, instruction-following, and agentic capabilities—all while prioritizing transparency and customization for the open-source community.
|
| 23 |
+
|
| 24 |
+
## Highlights
|
| 25 |
+
|
| 26 |
+
* **MoE Architecture (102B / 12B):** Built on a Mixture-of-Experts architecture with **102B total / 12B active parameters**. This design delivers the knowledge depth of a massive model with the inference speed and cost-efficiency of a much smaller model.
|
| 27 |
+
* **Massive Training Scale:** Pre-trained on **19.7 trillion tokens**, ensuring broad knowledge coverage and robust reasoning capabilities across various domains.
|
| 28 |
+
|
| 29 |
+
## Model Overview
|
| 30 |
+
|
| 31 |
+
* **Model Name:** Solar Open 100B
|
| 32 |
+
* **Hugging Face ID:** Upstage/Solar-Open-100B
|
| 33 |
+
* **Architecture:** Mixture-of-Experts (MoE)
|
| 34 |
+
* **Total Parameters:** 102.6B
|
| 35 |
+
* **Active Parameters:** 12B (per token)
|
| 36 |
+
* **Experts:** 129 Experts (top 8 among 128 Routed + 1 Shared)
|
| 37 |
+
* **Pre-training Tokens:** 19.7 Trillion
|
| 38 |
+
* **Context Length:** 128k
|
| 39 |
+
* **Training Hardware:** NVIDIA B200 GPUs
|
| 40 |
+
* **License:** **Solar-Apache License 2.0** (See [LICENSE](./LICENSE))
|
| 41 |
+
* **Hardware Requirements:**
|
| 42 |
+
* **Minimum:** 4x NVIDIA A100 (80GB)
|
| 43 |
+
|
| 44 |
+
## Performance
|
| 45 |
+
|
| 46 |
+
TBA
|
| 47 |
+
|
| 48 |
+
## Inference Quickstart
|
| 49 |
+
|
| 50 |
+
We recommend using the following generation parameters:
|
| 51 |
+
|
| 52 |
+
```
|
| 53 |
+
temperature=0.8
|
| 54 |
+
top_p=0.95
|
| 55 |
+
top_k=50
|
| 56 |
+
```
|
| 57 |
+
|
| 58 |
+
### Transformers
|
| 59 |
+
|
| 60 |
+
Install the required dependencies:
|
| 61 |
+
|
| 62 |
+
```bash
|
| 63 |
+
pip install -U transformers kernels torch accelerate
|
| 64 |
+
```
|
| 65 |
+
|
| 66 |
+
Run inference with the following code:
|
| 67 |
+
|
| 68 |
+
```python
|
| 69 |
+
import torch
|
| 70 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 71 |
+
|
| 72 |
+
MODEL_ID = "upstage/Solar-Open-100B"
|
| 73 |
+
|
| 74 |
+
# Load model and tokenizer
|
| 75 |
+
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
|
| 76 |
+
|
| 77 |
+
model = AutoModelForCausalLM.from_pretrained(
|
| 78 |
+
pretrained_model_name_or_path=MODEL_ID,
|
| 79 |
+
torch_dtype=torch.bfloat16,
|
| 80 |
+
device_map="auto",
|
| 81 |
+
trust_remote_code=True,
|
| 82 |
+
)
|
| 83 |
+
|
| 84 |
+
# Prepare input
|
| 85 |
+
messages = [{"role": "user", "content": "who are you?"}]
|
| 86 |
+
inputs = tokenizer.apply_chat_template(
|
| 87 |
+
messages,
|
| 88 |
+
tokenize=True,
|
| 89 |
+
add_generation_prompt=True,
|
| 90 |
+
return_dict=True,
|
| 91 |
+
return_tensors="pt",
|
| 92 |
+
)
|
| 93 |
+
inputs = inputs.to(model.device)
|
| 94 |
+
|
| 95 |
+
# Generate response
|
| 96 |
+
generated_ids = model.generate(
|
| 97 |
+
**inputs,
|
| 98 |
+
max_new_tokens=4096,
|
| 99 |
+
temperature=0.8,
|
| 100 |
+
top_p=0.95,
|
| 101 |
+
top_k=50,
|
| 102 |
+
do_sample=True,
|
| 103 |
+
)
|
| 104 |
+
generated_text = tokenizer.decode(generated_ids[0][inputs.input_ids.shape[1] :])
|
| 105 |
+
print(generated_text)
|
| 106 |
+
```
|
| 107 |
+
|
| 108 |
+
### vLLM
|
| 109 |
+
|
| 110 |
+
#### Option 1: Using Docker (Highly Recommended)
|
| 111 |
+
Docker is the **recommended deployment method** for running `Solar-Open-100B`.
|
| 112 |
+
|
| 113 |
+
```bash
|
| 114 |
+
# For 8 GPUs
|
| 115 |
+
docker run --gpus all \
|
| 116 |
+
--ipc=host \
|
| 117 |
+
-p 8000:8000 \
|
| 118 |
+
upstage/vllm-solar-open:latest \
|
| 119 |
+
upstage/Solar-Open-100B \
|
| 120 |
+
--trust-remote-code \
|
| 121 |
+
--enable-auto-tool-choice \
|
| 122 |
+
--tool-call-parser solar_open \
|
| 123 |
+
--reasoning-parser solar_open \
|
| 124 |
+
--logits-processors vllm.model_executor.models.parallel_tool_call_logits_processor:ParallelToolCallLogitsProcessor \
|
| 125 |
+
--logits-processors vllm.model_executor.models.solar_open_logits_processor:SolarOpenTemplateLogitsProcessor \
|
| 126 |
+
--tensor-parallel-size 8
|
| 127 |
+
```
|
| 128 |
+
|
| 129 |
+
#### Option 2: Installing from Source
|
| 130 |
+
For development, debugging, custom modifications or offline inference, Solar Open can also be run
|
| 131 |
+
using a source installation of vLLM. We recommend using **[uv](https://docs.astral.sh/uv/)** for environment
|
| 132 |
+
management and dependency resolution.
|
| 133 |
+
|
| 134 |
+
Create and activate a Python virtual environment
|
| 135 |
+
```bash
|
| 136 |
+
uv venv --python 3.12 --seed
|
| 137 |
+
source .venv/bin/activate
|
| 138 |
+
```
|
| 139 |
+
|
| 140 |
+
Install Solar Open's optimized vLLM
|
| 141 |
+
```bash
|
| 142 |
+
VLLM_PRECOMPILED_WHEEL_LOCATION="https://github.com/vllm-project/vllm/releases/download/v0.12.0/vllm-0.12.0-cp38-abi3-manylinux_2_31_x86_64.whl" \
|
| 143 |
+
VLLM_USE_PRECOMPILED=1 \
|
| 144 |
+
uv pip install git+https://github.com/UpstageAI/vllm.git@v0.12.0-solar-open
|
| 145 |
+
```
|
| 146 |
+
|
| 147 |
+
Start the vLLM server (For 8 GPUs)
|
| 148 |
+
```bash
|
| 149 |
+
vllm serve upstage/Solar-Open-100B \
|
| 150 |
+
--trust-remote-code \
|
| 151 |
+
--enable-auto-tool-choice \
|
| 152 |
+
--tool-call-parser solar_open \
|
| 153 |
+
--reasoning-parser solar_open \
|
| 154 |
+
--logits-processors vllm.model_executor.models.parallel_tool_call_logits_processor:ParallelToolCallLogitsProcessor \
|
| 155 |
+
--logits-processors vllm.model_executor.models.solar_open_logits_processor:SolarOpenTemplateLogitsProcessor \
|
| 156 |
+
--tensor-parallel-size 8
|
| 157 |
+
```
|
| 158 |
+
|
| 159 |
+
## Public API Access
|
| 160 |
+
|
| 161 |
+
The official API service for Solar Open is scheduled to launch publicly on **January**.
|
| 162 |
+
|
| 163 |
+
* **Access:** Upstage Console (TBA)
|
| 164 |
+
* **Documentation:** Upstage Console (TBA)
|
| 165 |
+
|
| 166 |
+
## Citation
|
| 167 |
+
|
| 168 |
+
If you use Solar Open in your research, please cite:
|
| 169 |
+
|
| 170 |
+
```bibtex
|
| 171 |
+
@misc{solar-open-2025,
|
| 172 |
+
title={Solar Open 100B},
|
| 173 |
+
author={Upstage AI},
|
| 174 |
+
year={2025},
|
| 175 |
+
url={https://huggingface.co/Upstage/Solar-Open-100B}
|
| 176 |
+
}
|
| 177 |
+
```
|
Solar-Open-100B.png
ADDED
|
generation_config.json
CHANGED
|
@@ -1,14 +1,14 @@
|
|
| 1 |
{
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
| 4 |
-
"do_sample": true,
|
| 5 |
"eos_token_id": [
|
| 6 |
2,
|
| 7 |
24,
|
| 8 |
25
|
| 9 |
],
|
| 10 |
"pad_token_id": 2,
|
|
|
|
|
|
|
| 11 |
"temperature": 0.8,
|
| 12 |
-
"top_p": 0.95
|
| 13 |
-
|
| 14 |
-
}
|
|
|
|
| 1 |
{
|
| 2 |
"_from_model_config": true,
|
| 3 |
"bos_token_id": 1,
|
|
|
|
| 4 |
"eos_token_id": [
|
| 5 |
2,
|
| 6 |
24,
|
| 7 |
25
|
| 8 |
],
|
| 9 |
"pad_token_id": 2,
|
| 10 |
+
"transformers_version": "4.57.3",
|
| 11 |
+
"do_sample": true,
|
| 12 |
"temperature": 0.8,
|
| 13 |
+
"top_p": 0.95
|
| 14 |
+
}
|
|
|
tokenizer_config.json
CHANGED
|
@@ -36751,14 +36751,12 @@
|
|
| 36751 |
"clean_up_tokenization_spaces": false,
|
| 36752 |
"eos_token": "<|endoftext|>",
|
| 36753 |
"extra_special_tokens": {},
|
| 36754 |
-
"model_input_names": [
|
| 36755 |
-
"input_ids",
|
| 36756 |
-
"attention_mask"
|
| 36757 |
-
],
|
| 36758 |
"model_max_length": 1000000000000000019884624838656,
|
| 36759 |
"pad_token": "<|endoftext|>",
|
| 36760 |
"padding_side": "left",
|
| 36761 |
"split_special_tokens": false,
|
| 36762 |
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 36763 |
-
"unk_token": "<unk>"
|
|
|
|
| 36764 |
}
|
|
|
|
| 36751 |
"clean_up_tokenization_spaces": false,
|
| 36752 |
"eos_token": "<|endoftext|>",
|
| 36753 |
"extra_special_tokens": {},
|
| 36754 |
+
"model_input_names": ["input_ids", "attention_mask"],
|
|
|
|
|
|
|
|
|
|
| 36755 |
"model_max_length": 1000000000000000019884624838656,
|
| 36756 |
"pad_token": "<|endoftext|>",
|
| 36757 |
"padding_side": "left",
|
| 36758 |
"split_special_tokens": false,
|
| 36759 |
"tokenizer_class": "PreTrainedTokenizerFast",
|
| 36760 |
+
"unk_token": "<unk>",
|
| 36761 |
+
"chat_template": "{#- ======== Template Parameters ======== #}\r\n{%- set add_generation_prompt = add_generation_prompt if add_generation_prompt is defined else true %}\r\n{%- set default_system_prompt = default_system_prompt if default_system_prompt is defined else true %}\r\n{%- set reasoning_effort = reasoning_effort if reasoning_effort is defined else \"high\" %}\r\n{%- set think_render_option = think_render_option if think_render_option is defined else \"lastthink\" %}\r\n\r\n{#- ======== System Block State ======== #}\r\n{%- set sys_ns = namespace(is_first_block=true) -%}\r\n\r\n{#- ======== Find last user message index ======== #}\r\n{%- set last_user_idx = namespace(value=-1) -%}\r\n{%- for message in messages -%}\r\n {%- if message.role == 'user' -%}\r\n {%- set last_user_idx.value = loop.index0 -%}\r\n {%- endif -%}\r\n{%- endfor -%}\r\n\r\n{#- ======== System messages renderers ======== #}\r\n{%- macro render_system_message(user_system_messages) %}\r\n {%- if default_system_prompt %}\r\n {%- if not sys_ns.is_first_block %}{{- \"\\n\\n\" }}{%- endif %}\r\n {%- set sys_ns.is_first_block = false %}\r\n {{- \"## Provider System Prompt\\n\\nYou are Solar Open 100B, a large language model trained by Upstage AI, a Korean startup. Your knowledge cutoff is 2025-07. The current date is \" + strftime_now(\"%Y-%m-%d\") + \".\" }}\r\n {%- endif -%}\r\n {%- if user_system_messages %}\r\n {%- if not sys_ns.is_first_block %}{{- \"\\n\\n\" }}{%- endif %}\r\n {%- set sys_ns.is_first_block = false %}\r\n {{- \"## System Prompt\" }}\r\n {%- for system_message in user_system_messages %}\r\n {{- \"\\n\\n\" }}\r\n {{- system_message }}\r\n {%- endfor %}\r\n {%- endif -%}\r\n{%- endmacro %}\r\n\r\n{%- macro render_tool_instruction(tools) %}\r\n {%- if not sys_ns.is_first_block %}{{- \"\\n\\n\" }}{%- endif %}\r\n {%- set sys_ns.is_first_block = false %}\r\n {{- \"## Tools\\n\\n### Tool Call Instruction\" }}\r\n {{- \"\\nYou may invoke one or more tools to assist with the user's query. Available tools are provided in JSON Schema format: <|tools:begin|><|tool:begin|><tools-json-object><|tool:end|>...<|tools:end|>\\n\" }}\r\n {{- \"\\n### Available Tools\\n\" }}\r\n {{- \"<|tools:begin|>\" }}\r\n {%- for tool in tools %}\r\n {{- \"<|tool:begin|>\" }}\r\n {{- tool.function | tojson }}\r\n {{- \"<|tool:end|>\" }}\r\n {%- endfor %}\r\n {{- \"<|tools:end|>\\n\" }}\r\n {{- \"\\n### Tool Call Format\\n\" }}\r\n {{- \"For each tool call, return a JSON object with the following structure, enclosed within <|tool_call:begin|> and <|tool_call:end|> tags: \\n<|tool_call:begin|><tool-call-id><|tool_call:name|><tool-name><|tool_call:args|><args-json-object><|tool_call:end|>\\n\" }}\r\n {{- \"- The <tool-call-id> must be a randomly generated string consisting of 10 lowercase letters (a-z) and\/or digits (0-9) (e.g., a1b2c3d4e5)\\n\" }}\r\n {{- \"\\n### Tool Response Format\\n\" }}\r\n {{- \"Each tool is responded by `tool` with the following structure:\\n<|tool_response:id|><tool-call-id><|tool_response:name|><tool-name><|tool_response:result|><results><|tool_response:end|>\\n\" }}\r\n {{- \"- Ensure the <tool-call-id> matches the corresponding tool call\" -}}\r\n{%- endmacro %}\r\n\r\n{%- macro render_json_response_format_instruction(response_format) %}\r\n {%- if not sys_ns.is_first_block %}{{- \"\\n\\n\" }}{%- endif %}\r\n {%- set sys_ns.is_first_block = false %}\r\n {{- \"## Output Format Constraint\" }}\r\n {{- \"\\n\\nYour final response should follow the JSON schema: \\n[Start of schema]\" }}\r\n {{- response_format }}\r\n {{- \"\\n[End of schema]\\nPlease ensure your answers adhere to this format and do not contain any unnecessary text.\" }}\r\n{%- endmacro %}\r\n\r\n{%- macro get_tool_name(messages, tool_call_id) %}\r\n {%- for msg in messages -%}\r\n {%- if msg.role == 'assistant' and msg.tool_calls -%}\r\n {%- for tool_call in msg.tool_calls -%}\r\n {%- if tool_call.id == tool_call_id -%}\r\n {{- tool_call.function.name }}\r\n {%- endif -%}\r\n {%- endfor -%}\r\n {%- endif -%}\r\n {%- endfor -%}\r\n{%- endmacro %}\r\n\r\n{%- macro render_tool_arguments(tool_arguments) %}\r\n {%- if tool_arguments is mapping -%}\r\n {{- tool_arguments | tojson }}\r\n {%- else -%}\r\n {{- tool_arguments }}\r\n {%- endif -%}\r\n{%- endmacro %}\r\n\r\n{#- ======== Render system message ======== #}\r\n{%- set ns = namespace(system_messages=[]) -%}\r\n{%- for message in messages -%}\r\n {%- if message.role == 'system' -%}\r\n {%- set ns.system_messages = ns.system_messages + [message.content] -%}\r\n {%- endif -%}\r\n{%- endfor -%}\r\n\r\n{%- if ns.system_messages or default_system_prompt or tools or response_format -%}\r\n {{- \"<|begin|>system<|content|>\" }}\r\n {{- render_system_message(ns.system_messages) }}\r\n {%- if tools -%}\r\n {{- render_tool_instruction(tools) }}\r\n {%- endif %}\r\n {%- if response_format -%}\r\n {{- render_json_response_format_instruction(response_format) }}\r\n {%- endif %}\r\n {{- \"<|end|>\" }}\r\n{%- endif -%}\r\n\r\n{#- ======== Render main messages ======== #}\r\n{%- for message in messages -%}\r\n {%- if message.role == 'user' -%}\r\n {{- \"<|begin|>user<|content|>\" + message.content + \"<|end|>\" }}\r\n {%- elif message.role == 'tool' -%}\r\n {%- set prev_is_tool = loop.index0 > 0 and messages[loop.index0 - 1].role == 'tool' -%}\r\n {%- set next_is_tool = loop.index0 < (messages | length - 1) and messages[loop.index0 + 1].role == 'tool' -%}\r\n {%- if not prev_is_tool -%}\r\n {{- \"<|begin|>tool<|tool_response|>\" }}\r\n {%- endif -%}\r\n {{- \"<|tool_response:begin|>\" + message.tool_call_id + \"<|tool_response:name|>\" }}\r\n {{- get_tool_name(messages, message.tool_call_id) }}\r\n {{- \"<|tool_response:result|>\" }}\r\n {{- message.content }}\r\n {{- \"<|tool_response:end|>\" }}\r\n {%- if not next_is_tool -%}\r\n {{- \"<|end|>\" }}\r\n {%- endif -%}\r\n {%- elif message.role == 'assistant' -%}\r\n {#- ======== Assistant Thinking ======== #}\r\n {%- if think_render_option == \"all\" -%}\r\n {%- if message.reasoning -%}\r\n {{- \"<|begin|>assistant<|think|>\" + message.reasoning + \"<|end|>\" }}\r\n {%- endif -%}\r\n {%- elif think_render_option == \"lastthink\" -%}\r\n {%- if message.reasoning and loop.index0 > last_user_idx.value -%}\r\n {{- \"<|begin|>assistant<|think|>\" + message.reasoning + \"<|end|>\" }}\r\n {%- endif -%}\r\n {%- endif -%}\r\n\r\n {#- ======== Assistant Messages ======== #}\r\n {%- if message.tool_calls -%}\r\n {{- \"<|begin|>assistant<|tool_calls|>\" }}\r\n {%- for tool_call in message.tool_calls -%}\r\n {{- \"<|tool_call:begin|>\" + tool_call.id +\"<|tool_call:name|>\" + tool_call.function.name + \"<|tool_call:args|>\" }}\r\n {{- render_tool_arguments(tool_call.function.arguments) }}\r\n {{- \"<|tool_call:end|>\" }}\r\n {%- endfor -%}\r\n {{- \"<|calls|>\" }}\r\n {%- else -%}\r\n {{- \"<|begin|>assistant<|content|>\" + message.content + \"<|end|>\" }}\r\n {%- endif -%}\r\n {%- endif -%}\r\n{%- endfor -%}\r\n\r\n{%- if add_generation_prompt -%}\r\n {%- if reasoning_effort in [\"low\", \"minimal\"] -%}\r\n {{- \"<|begin|>assistant<|think|><|end|>\" }}\r\n {%- endif -%}\r\n {{- \"<|begin|>assistant\" }}\r\n{%- endif -%}"
|
| 36762 |
}
|