Text Generation
Safetensors
English
fleck
causal-language-model
custom-architecture
conversational
research
brulee-1 commited on
Commit
03f87b0
·
verified ·
1 Parent(s): 215dc3b

fix: publish HF model name and links

Browse files
Files changed (4) hide show
  1. README.md +3 -3
  2. config.json +1 -1
  3. inference.py +1 -1
  4. model.safetensors +2 -2
README.md CHANGED
@@ -13,9 +13,9 @@ tags:
13
  base_model: ML-is-Fun/Fleck-S-100K-Base
14
  ---
15
 
16
- # Fleck-S-100K
17
 
18
- **Fleck-S-B-factorized-wide Instruct** — the independently instruction-tuned child of [`Fleck-S-100K-Base`](https://huggingface.co/ML-is-Fun/Fleck-S-100K-Base).
19
 
20
  - **109,384 parameters**
21
  - **BF16 SafeTensors weights**
@@ -50,7 +50,7 @@ The Base parent was initialized from scratch and pretrained for exactly 100,000,
50
 
51
  | | |
52
  | --- | --- |
53
- | Base model | `Fleck-S-100K-Base` |
54
  | Dataset | [`databricks/databricks-dolly-15k`](https://huggingface.co/datasets/databricks/databricks-dolly-15k) |
55
  | Dataset license | CC BY-SA 3.0 |
56
  | Requested budget | 100K supervised tokens |
 
13
  base_model: ML-is-Fun/Fleck-S-100K-Base
14
  ---
15
 
16
+ # [Fleck-S-100K](https://huggingface.co/ML-is-Fun/Fleck-S-100K)
17
 
18
+ **[Fleck-S-100K Instruct](https://huggingface.co/ML-is-Fun/Fleck-S-100K)** — the independently instruction-tuned child of [`Fleck-S-100K-Base`](https://huggingface.co/ML-is-Fun/Fleck-S-100K-Base).
19
 
20
  - **109,384 parameters**
21
  - **BF16 SafeTensors weights**
 
50
 
51
  | | |
52
  | --- | --- |
53
+ | Base model | [`Fleck-S-100K-Base`](https://huggingface.co/ML-is-Fun/Fleck-S-100K-Base) |
54
  | Dataset | [`databricks/databricks-dolly-15k`](https://huggingface.co/datasets/databricks/databricks-dolly-15k) |
55
  | Dataset license | CC BY-SA 3.0 |
56
  | Requested budget | 100K supervised tokens |
config.json CHANGED
@@ -3,7 +3,7 @@
3
  "model_type": "fleck",
4
  "custom_architecture": true,
5
  "implementation": "inference.py",
6
- "candidate_id": "Fleck-S-B-factorized-wide",
7
  "vocab_size": 1024,
8
  "hidden_size": 64,
9
  "intermediate_size": 128,
 
3
  "model_type": "fleck",
4
  "custom_architecture": true,
5
  "implementation": "inference.py",
6
+ "candidate_id": "Fleck-S-100K",
7
  "vocab_size": 1024,
8
  "hidden_size": 64,
9
  "intermediate_size": 128,
inference.py CHANGED
@@ -36,7 +36,7 @@ SYSTEM_ID, USER_ID, ASSISTANT_ID, EOT_ID = 4, 5, 6, 7
36
  EXECUTION_ORDER = (0, 1, 0, 1)
37
  TOKENIZER_NAME = "Fleck-Tokenizer-1024"
38
  MODEL_METADATA = {
39
- "candidate_id": "Fleck-S-B-factorized-wide",
40
  "canonical_dtype": "bfloat16",
41
  "context_length": "2048",
42
  "effective_depth": "4",
 
36
  EXECUTION_ORDER = (0, 1, 0, 1)
37
  TOKENIZER_NAME = "Fleck-Tokenizer-1024"
38
  MODEL_METADATA = {
39
+ "candidate_id": "Fleck-S-100K",
40
  "canonical_dtype": "bfloat16",
41
  "context_length": "2048",
42
  "effective_depth": "4",
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ca2ab46b51ea1cc6d7ae832ba3c68ee4348a023d72418f09500e20016d19c3a9
3
- size 222216
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9d47bae7c83dab36d7d3e1aca47b09ee97ad59810ccedf58743f30a69151422
3
+ size 222208