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
Not-For-All-Audiences
Inference Endpoints
Update README.md
Browse files
README.md
CHANGED
|
@@ -257,6 +257,12 @@ This is due to "random" pruning method in "DARE-TIES".
|
|
| 257 |
|
| 258 |
Mistral Nemo models used here seem acutely sensitive to this process.
|
| 259 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 260 |
"tokenizer_source: union" is used so that multiple "templates" work and each fine tune uses one or two of the templates.
|
| 261 |
|
| 262 |
<h3>EXAMPLES PROMPTS and OUTPUT:</h3>
|
|
|
|
| 257 |
|
| 258 |
Mistral Nemo models used here seem acutely sensitive to this process.
|
| 259 |
|
| 260 |
+
This shows up in PPL and "real world" tests as well as "TEMP=0" ("core test") generations.
|
| 261 |
+
|
| 262 |
+
PPL range of 7.7327 to 7.8024 ... and that is on just 10 generations.
|
| 263 |
+
|
| 264 |
+
This model: PPL = 7.7729 +/- 0.12700 (100 chunks, wiki.test.raw)
|
| 265 |
+
|
| 266 |
"tokenizer_source: union" is used so that multiple "templates" work and each fine tune uses one or two of the templates.
|
| 267 |
|
| 268 |
<h3>EXAMPLES PROMPTS and OUTPUT:</h3>
|