Update README.md

README.md

@@ -36,6 +36,32 @@ Note that tensor parallelism is not currently supported for this architecture, s
 [measurement.json - 3.0bpw_H6 vs 4.0bpw_H6](https://huggingface.co/MikeRoz/MiniMax-M2.5-exl3/blob/main/measurement_MiniMaxAI_MiniMax-M2.5-3.0-4.0.json)    
 [measurement.json - 4.0bpw_H6 vs 5.0bpw_H6](https://huggingface.co/MikeRoz/MiniMax-M2.5-exl3/blob/main/measurement_MiniMaxAI_MiniMax-M2.5-4.0-5.0.json)    
 
+### matplotlib Catbench
+
+<details><summary>Click to see cat plots!</summary>
+
+2.0 bpw:
+![2.0 Catbench](minimax_m2.5_matplotlib_catbench_2.0bpw.png)
+2.1 bpw:
+![2.1 Catbench](minimax_m2.5_matplotlib_catbench_2.1bpw.png)
+2.5 bpw:
+![2.5 Catbench](minimax_m2.5_matplotlib_catbench_2.5bpw.png)
+3.0 bpw:
+![3.0 Catbench](minimax_m2.5_matplotlib_catbench_3.0bpw.png)
+3.06 bpw:
+![3.06 Catbench](minimax_m2.5_matplotlib_catbench_3.06bpw.png)
+3.5 bpw:
+![3.5 Catbench](minimax_m2.5_matplotlib_catbench_3.5bpw.png)
+4.0 bpw:
+![4.0 Catbench](minimax_m2.5_matplotlib_catbench_4.0bpw.png)
+5.0 bpw:
+![5.0 Catbench](minimax_m2.5_matplotlib_catbench_5.0bpw.png)
+
+Prompted in Text Generation Web UI in chat-instruct mode with MiniMax AI's recommended settings (temp 1, top p 0.95, top k 40).
+
+Note that 2.1, 4.0, and 5.0 all required a single re-roll each to get a working script.
+</details>
+
 ### How to use these quants
 
 The documentation for [exllamav3](https://github.com/turboderp-org/exllamav3/) is your best bet here, as well as that of [TabbyAPI](https://github.com/theroyallab/tabbyAPI) or [Text Generation Web UI (oobabooga)](https://github.com/oobabooga/text-generation-webui). In short: