Move Lambda support section to card footer

README.md

@@ -23,28 +23,6 @@ tags:
 
 # lambda-1-360m-mid
 
-## Support Lambda
-
-[Lambda](https://github.com/KeisukeMiyamoto1324/lambda) is a project developed by a student to build small Japanese language models from scratch. The datasets, source code, and model weights are publicly available.
-
-Training language models is very expensive, and I have funded the project with income from my part-time job. As the models grow, the cost is becoming too high for me to maintain the current development pace.
-
-You can support Lambda in two ways: use the Vast.ai referral link when purchasing GPU credits or make a donation through Ko-fi. Thank you for helping Lambda continue to grow.
-
-### PR: vast.ai
-
-Vast.ai is a GPU cloud platform that lets you rent powerful GPUs from providers around the world. It is often more affordable than major cloud providers, with **NVIDIA H100 SXM GPUs available from around $1.54 per hour**. Check out Vast.ai for a lower-cost way to train language models.
-
-When you sign up and purchase credits through the link below, I receive GPU credits equal to 3% of your purchase. This does not reduce your credits or add any extra cost. Using this link directly supports the development of Lambda.
-
-Link: https://cloud.vast.ai/?ref_id=521936
-
-### Donate
-
-You can support the project with a donation starting from $5 through Ko-fi. Your support helps cover GPU costs and allows me to continue developing larger models.
-
-Link: https://ko-fi.com/lambda_llm
-
 lambda-1-360m-mid is an experimental Japanese language model based on [`lambda-1-360m-base`](https://huggingface.co/KeisukeMiyamoto/lambda-1-360m-base) and further pretrained on a synthetic Japanese textbook corpus.
 
 All training code is publicly available at [KeisukeMiyamoto1324/lambda](https://github.com/KeisukeMiyamoto1324/lambda).
@@ -90,3 +68,27 @@ python3 src/inference_base/inference_hf.py \
 This model is not instruction-tuned or safety-aligned. It may generate incorrect, biased, unsafe, or low-quality text.
 
 The model was trained primarily on Japanese text and has not been evaluated on standard benchmarks.
+
+---
+
+## Support Lambda
+
+[Lambda](https://github.com/KeisukeMiyamoto1324/lambda) is a project developed by a student to build small Japanese language models from scratch. The datasets, source code, and model weights are publicly available.
+
+Training language models is very expensive, and I have funded the project with income from my part-time job. As the models grow, the cost is becoming too high for me to maintain the current development pace.
+
+You can support Lambda in two ways: use the Vast.ai referral link when purchasing GPU credits or make a donation through Ko-fi. Thank you for helping Lambda continue to grow.
+
+### PR: vast.ai
+
+Vast.ai is a GPU cloud platform that lets you rent powerful GPUs from providers around the world. It is often more affordable than major cloud providers, with **NVIDIA H100 SXM GPUs available from around $1.54 per hour**. Check out Vast.ai for a lower-cost way to train language models.
+
+When you sign up and purchase credits through the link below, I receive GPU credits equal to 3% of your purchase. This does not reduce your credits or add any extra cost. Using this link directly supports the development of Lambda.
+
+Link: https://cloud.vast.ai/?ref_id=521936
+
+### Donate
+
+You can support the project with a donation starting from $5 through Ko-fi. Your support helps cover GPU costs and allows me to continue developing larger models.
+
+Link: https://ko-fi.com/lambda_llm