tg-techie-agents commited on
Commit
fc0be89
·
verified ·
1 Parent(s): a89c17e

latest.json: checkpoints/4475515k (step 117,600)

Browse files
Files changed (4) hide show
  1. README.md +4 -4
  2. latest.json +7 -7
  3. latest/config.json +2 -2
  4. latest/model.safetensors +1 -1
README.md CHANGED
@@ -30,7 +30,7 @@ deliberately derived from it: Nirca began as a port of mini-AGI to MLX.
30
  ➡️ **Try it in your browser (WebGPU):** https://tg-techie-agents.github.io/nirca-mini-webgpu/mini/
31
 
32
  This card is written by program from the checkpoint's own record, at step
33
- 117,496 (4,472,825,935 tokens seen).
34
 
35
  ## Model details
36
 
@@ -115,9 +115,9 @@ Each step is listed because the run's records show it was done:
115
 
116
  ## Training procedure
117
 
118
- - **Steps:** 117,496
119
- - **Tokens seen:** 4,472,825,935 (one token is one byte)
120
- - **Quantization-aware training:** from step 14,655 of 117,496, so for 99% of its 4,472,825,935 tokens
121
  - **Objective:** next-byte prediction; on chats, loss on the assistant's turns only
122
  - **Recurrence:** the number of passes is drawn per step during training, and
123
  the halting head learns when to stop
 
30
  ➡️ **Try it in your browser (WebGPU):** https://tg-techie-agents.github.io/nirca-mini-webgpu/mini/
31
 
32
  This card is written by program from the checkpoint's own record, at step
33
+ 117,600 (4,475,515,245 tokens seen).
34
 
35
  ## Model details
36
 
 
115
 
116
  ## Training procedure
117
 
118
+ - **Steps:** 117,600
119
+ - **Tokens seen:** 4,475,515,245 (one token is one byte)
120
+ - **Quantization-aware training:** from step 14,655 of 117,600, so for 99% of its 4,475,515,245 tokens
121
  - **Objective:** next-byte prediction; on chats, loss on the assistant's turns only
122
  - **Recurrence:** the number of passes is drawn per step during training, and
123
  the halting head learns when to stop
latest.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "name": "nirca-mini",
3
- "step": 117496,
4
  "quant": "ternary",
5
  "quant_scope": "all",
6
  "n_experts": 32,
@@ -40,17 +40,17 @@
40
  "pool_depth": 1,
41
  "pool_max": 32
42
  },
43
- "tokens": 4472825935,
44
- "checkpoint": "checkpoints/4472825k/",
45
  "files": {
46
  "config.json": {
47
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4472825k/config.json",
48
- "sha256": "040a86d12806b4ea1fc65428979940f963b8b8146da5a3e6f8c9e1331433cb09",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4472825k/model.safetensors",
53
- "sha256": "6c9de3641f2512157f91c22c4d8c4ef6dcdf36c75464229adc7fc254722ee656",
54
  "bytes": 32638048
55
  }
56
  }
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 117600,
4
  "quant": "ternary",
5
  "quant_scope": "all",
6
  "n_experts": 32,
 
40
  "pool_depth": 1,
41
  "pool_max": 32
42
  },
43
+ "tokens": 4475515245,
44
+ "checkpoint": "checkpoints/4475515k/",
45
  "files": {
46
  "config.json": {
47
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4475515k/config.json",
48
+ "sha256": "9e0059157abe57f91f7cb84463c41202d005fe695e695455becd6282d82b9f50",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4475515k/model.safetensors",
53
+ "sha256": "176b52dbd98ae6d283ca25b89cbf4413799092c04ca20a57077d617d433970dd",
54
  "bytes": 32638048
55
  }
56
  }
latest/config.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "name": "nirca-mini",
3
- "step": 117496,
4
- "tokens": 4472825935,
5
  "quant": "ternary",
6
  "quant_scope": "all",
7
  "n_experts": 32,
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 117600,
4
+ "tokens": 4475515245,
5
  "quant": "ternary",
6
  "quant_scope": "all",
7
  "n_experts": 32,
latest/model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:6c9de3641f2512157f91c22c4d8c4ef6dcdf36c75464229adc7fc254722ee656
3
  size 32638048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:176b52dbd98ae6d283ca25b89cbf4413799092c04ca20a57077d617d433970dd
3
  size 32638048