Update README.md
Browse files
README.md
CHANGED
@@ -27,7 +27,6 @@ datasets:
|
|
27 |
## Model Overview
|
28 |
- **Name**: `aquif-neo`
|
29 |
- **Parameters**: 64.1 million
|
30 |
-
- **Context Window**: 128,000 tokens
|
31 |
- **Architecture**: Dense
|
32 |
- **Type**: General-purpose LLM
|
33 |
- **Hosted on**: [Hugging Face](https://huggingface.co/aquiffoo/aquif-neo)
|
@@ -35,18 +34,33 @@ datasets:
|
|
35 |
## Training Steps
|
36 |
|
37 |
step 500 | loss = 0.9147
|
|
|
38 |
step 1000 | loss = 0.7440
|
|
|
39 |
step 1500 | loss = 0.6791
|
|
|
40 |
step 2000 | loss = 0.6631
|
|
|
41 |
step 2500 | loss = 0.6439
|
|
|
42 |
step 3000 | loss = 0.6335
|
|
|
43 |
step 3500 | loss = 0.6176
|
|
|
44 |
step 4000 | loss = 0.5987
|
|
|
45 |
step 4500 | loss = 0.5979
|
|
|
46 |
step 5000 | loss = 0.6018
|
|
|
47 |
step 5500 | loss = 0.5767
|
|
|
48 |
step 6000 | loss = 0.5839
|
|
|
49 |
step 6500 | loss = 0.5754
|
|
|
50 |
step 7000 | loss = 0.5644
|
|
|
51 |
step 7500 | loss = 0.5640
|
|
|
52 |
step 8000 | loss = 0.5686
|
|
|
27 |
## Model Overview
|
28 |
- **Name**: `aquif-neo`
|
29 |
- **Parameters**: 64.1 million
|
|
|
30 |
- **Architecture**: Dense
|
31 |
- **Type**: General-purpose LLM
|
32 |
- **Hosted on**: [Hugging Face](https://huggingface.co/aquiffoo/aquif-neo)
|
|
|
34 |
## Training Steps
|
35 |
|
36 |
step 500 | loss = 0.9147
|
37 |
+
\
|
38 |
step 1000 | loss = 0.7440
|
39 |
+
\
|
40 |
step 1500 | loss = 0.6791
|
41 |
+
\
|
42 |
step 2000 | loss = 0.6631
|
43 |
+
\
|
44 |
step 2500 | loss = 0.6439
|
45 |
+
\
|
46 |
step 3000 | loss = 0.6335
|
47 |
+
\
|
48 |
step 3500 | loss = 0.6176
|
49 |
+
\
|
50 |
step 4000 | loss = 0.5987
|
51 |
+
\
|
52 |
step 4500 | loss = 0.5979
|
53 |
+
\
|
54 |
step 5000 | loss = 0.6018
|
55 |
+
\
|
56 |
step 5500 | loss = 0.5767
|
57 |
+
\
|
58 |
step 6000 | loss = 0.5839
|
59 |
+
\
|
60 |
step 6500 | loss = 0.5754
|
61 |
+
\
|
62 |
step 7000 | loss = 0.5644
|
63 |
+
\
|
64 |
step 7500 | loss = 0.5640
|
65 |
+
\
|
66 |
step 8000 | loss = 0.5686
|