Text Generation
GGUF
English
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 prosing
vivid writing
fiction
roleplaying
bfloat16
swearing
rp
horror
mistral nemo
mergekit
Inference Endpoints
Update README.md
Browse files
README.md
CHANGED
|
@@ -261,6 +261,12 @@ This is due to "random" pruning method in "DARE-TIES".
|
|
| 261 |
|
| 262 |
Mistral Nemo models used here seem acutely sensitive to this process.
|
| 263 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 264 |
"tokenizer_source: union" is used so that multiple "templates" work and each fine tune uses one or two of the templates.
|
| 265 |
|
| 266 |
<h3>EXAMPLES PROMPTS and OUTPUT:</h3>
|
|
|
|
| 261 |
|
| 262 |
Mistral Nemo models used here seem acutely sensitive to this process.
|
| 263 |
|
| 264 |
+
This shows up in PPL and "real world" tests as well as "TEMP=0" ("core test") generations.
|
| 265 |
+
|
| 266 |
+
PPL range of 7.7327 to 7.8024 ... and that is on just 10 generations.
|
| 267 |
+
|
| 268 |
+
This model: PPL = 7.7327 +/- 0.12567 (100 chunks, wiki.test.raw)
|
| 269 |
+
|
| 270 |
"tokenizer_source: union" is used so that multiple "templates" work and each fine tune uses one or two of the templates.
|
| 271 |
|
| 272 |
<h3>EXAMPLES PROMPTS and OUTPUT:</h3>
|