NovatasticRoScript commited on
Commit
d6eb4a3
·
verified ·
1 Parent(s): 7e86017

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -2
README.md CHANGED
@@ -13,14 +13,14 @@ tags:
13
  pipeline_tag: text-generation
14
  ---
15
 
16
- # Atomight-V2.5-1.7B
17
 
18
  <p align="center">
19
  <img src="OfficialAtomight.png" alt="Atomight Logo" width="500" style="max-width: 100%;">
20
  </p>
21
 
22
  **Atomight V2.5 1.7B**, our *most capable and intelligent model* yet, is a multi-domain reasoning model built on **Qwen3-1.7B**, trained with **GRPO** (Group Relative Policy Optimization) across STEM, science, math, and code, using a curated six-dataset stack in a streaming round-robin setup — trained entirely on a single Colab free-tier T4 GPU.
23
-
24
  ## Overview
25
 
26
  Atomight-V2.5 is post-trained from `Qwen/Qwen3-1.7B` (already instruct/chat-capable) using memory-efficient chunked GRPO to fit T4 VRAM constraints. Training data spans four domains, each selected for having genuinely verifiable rewards (exact-match answers or executable test cases) rather than relying purely on LLM-judged quality:
 
13
  pipeline_tag: text-generation
14
  ---
15
 
16
+ # Atomight-V2.5-1.7B-C1
17
 
18
  <p align="center">
19
  <img src="OfficialAtomight.png" alt="Atomight Logo" width="500" style="max-width: 100%;">
20
  </p>
21
 
22
  **Atomight V2.5 1.7B**, our *most capable and intelligent model* yet, is a multi-domain reasoning model built on **Qwen3-1.7B**, trained with **GRPO** (Group Relative Policy Optimization) across STEM, science, math, and code, using a curated six-dataset stack in a streaming round-robin setup — trained entirely on a single Colab free-tier T4 GPU.
23
+ **Note: This is only the first checkpoint model of Atomight-V2.5-1.7B, but still usable and powerful model.**
24
  ## Overview
25
 
26
  Atomight-V2.5 is post-trained from `Qwen/Qwen3-1.7B` (already instruct/chat-capable) using memory-efficient chunked GRPO to fit T4 VRAM constraints. Training data spans four domains, each selected for having genuinely verifiable rewards (exact-match answers or executable test cases) rather than relying purely on LLM-judged quality: