remove hfviewer embed (refuses custom-code models); keep Hub frontmatter

README.md

@@ -23,7 +23,7 @@ datasets:
 
 # min-spark
 
-**min-spark is a 5.76M-parameter language model with native effort levels.** It introduces controllable depth to the sub-10M model class with the use of native effort levels.
+**min-spark is a 5.76M-parameter language model with native effort levels.** It introduces controllable depth to the sub-10M model class. The same checkpoint can produce a quick completion or spend more computation on the same prompt, selected with one inference argument.
 
 The model was trained on 10.01B tokens. Its looped decoder reuses a compact transformer core across multiple passes, giving a small model the computation of a deeper network while keeping the parameter count fixed.
 
@@ -137,10 +137,6 @@ KeyLM reports 29.9% on its ARC average, 29.7% on HellaSwag, and 60.0% on PIQA. m
 | Context window | 512 tokens |
 | Effort (loop count) | low = 2 · medium = 3 · high = 4 |
 
-[![min-spark architecture](https://hfviewer.com/api/card.svg?source=MinimaLabs%2Fmin-spark&granularity=auto&animated=false)](https://hfviewer.com/MinimaLabs/min-spark)
-
-Open the image to explore the full architecture graph in hfviewer, including the block-level structure and parameter flow.
-
 ## Training
 
 | Field | Value |
@@ -176,4 +172,4 @@ The model supports a 512-token context window. Generation accepts one sequence a
 
 ## License
 
-Apache-2.0. See [LICENSE](LICENSE).
+Apache-2.0. See [LICENSE](LICENSE).