teszenofficial commited on
Commit
64eb002
·
verified ·
1 Parent(s): bb32bd8

Upload 3 files

Browse files
Files changed (3) hide show
  1. MTP1_MODEL.pt +3 -0
  2. MTP1_config.json +37 -0
  3. MTP1_tokenizer.model +3 -0
MTP1_MODEL.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c6c208900ebccbd76d28b7c8f1aabfc5d41778845cf399deeb07fe02e387805
3
+ size 142548993
MTP1_config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "vocab_size": 8000,
3
+ "block_size": 384,
4
+ "n_layer": 10,
5
+ "n_head": 8,
6
+ "n_embd": 512,
7
+ "dropout": 0.2,
8
+ "tokenizer_type": "sentencepiece_bpe",
9
+ "tokenizer_file": "MTP1_tokenizer.model",
10
+ "special_tokens": {
11
+ "pad_id": 0,
12
+ "bos_id": 1,
13
+ "eos_id": 2,
14
+ "unk_id": 3
15
+ },
16
+ "generation_defaults": {
17
+ "max_new_tokens": 200,
18
+ "temperature": 1.05,
19
+ "top_k": 40,
20
+ "top_p": 0.95,
21
+ "repetition_penalty": 1.2,
22
+ "no_repeat_ngram_size": 3
23
+ },
24
+ "meta": {
25
+ "model_name": "MTP-1",
26
+ "architecture": "gpt_rope_swiglu_rmsnorm",
27
+ "trained_examples": 20495,
28
+ "augmented_examples": 58871,
29
+ "corpus_fragments": 573,
30
+ "total_sequences": 61489,
31
+ "val_sequences": 2102,
32
+ "best_val_loss": 4.375298156738281,
33
+ "stopped_early": true,
34
+ "label_smoothing": 0.1,
35
+ "created": "2026-08-14 22:19:24"
36
+ }
37
+ }
MTP1_tokenizer.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2140d1494be1fc30b749059ecc4a4195e325c2060bed066ff7a827a93dbe91f
3
+ size 371988