Text Generation
GGUF
Magistral-Small-2506
bfloat16
thinking
reasoning
all use cases
creative use cases
creative
creative writing
fiction writing
plot generation
sub-plot generation
story generation
scene continue
storytelling
fiction story
science fiction
romance
all genres
story
writing
vivid prose
vivid writing
fiction
roleplaying
swearing
rp
NEO Imatrix
GGUFs
Maxed Output Tensor
imatrix
conversational
Update README.md
Browse files
README.md
CHANGED
|
@@ -70,13 +70,21 @@ of Mistral's new reasoning model "Magistral-Small-2506":
|
|
| 70 |
|
| 71 |
https://huggingface.co/mistralai/Magistral-Small-2506/
|
| 72 |
|
| 73 |
-
|
| 74 |
- Quanted using NEO Imatrix Dataset
|
| 75 |
- The Output Tensor is set at BF16 / 16 bit full precision.
|
| 76 |
- Correct Jijna template which includes "System Prompt" embedded for reasoning.
|
| 77 |
-
- 32K / 32,768 context max
|
|
|
|
| 78 |
|
| 79 |
-
An additional repo of GGUFs set at 128k / 131,072 context will follow
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
Special thanks to "MLX-Community" for correct config/tokenizer files.
|
| 82 |
|
|
|
|
| 70 |
|
| 71 |
https://huggingface.co/mistralai/Magistral-Small-2506/
|
| 72 |
|
| 73 |
+
About these GGUFS:
|
| 74 |
- Quanted using NEO Imatrix Dataset
|
| 75 |
- The Output Tensor is set at BF16 / 16 bit full precision.
|
| 76 |
- Correct Jijna template which includes "System Prompt" embedded for reasoning.
|
| 77 |
+
- 32K / 32,768 context max (default/set at org repo)
|
| 78 |
+
- Suggest min context of 4k-8k for reasoning/output.-
|
| 79 |
|
| 80 |
+
An additional repo of GGUFs set at 128k / 131,072 context will follow, as per Mistrals notes that the model
|
| 81 |
+
was trained at 128k max context.
|
| 82 |
+
|
| 83 |
+
Please see notes at:
|
| 84 |
+
|
| 85 |
+
https://huggingface.co/mistralai/Magistral-Small-2506/
|
| 86 |
+
|
| 87 |
+
For temp, topk, top p and other suggested parameter settings.
|
| 88 |
|
| 89 |
Special thanks to "MLX-Community" for correct config/tokenizer files.
|
| 90 |
|