tg-techie-agents commited on
Commit
6ca77db
·
verified ·
1 Parent(s): be342a2

latest.json: checkpoints/4599934k (step 122,510)

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
- 122,306 (4,595,055,325 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:** 122,306
119
- - **Tokens seen:** 4,595,055,325 (one token is one byte)
120
- - **Quantization-aware training:** from step 14,655 of 122,306, so for 99% of its 4,595,055,325 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
+ 122,510 (4,599,934,752 tokens seen).
34
 
35
  ## Model details
36
 
 
115
 
116
  ## Training procedure
117
 
118
+ - **Steps:** 122,510
119
+ - **Tokens seen:** 4,599,934,752 (one token is one byte)
120
+ - **Quantization-aware training:** from step 14,655 of 122,510, so for 99% of its 4,599,934,752 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": 122306,
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": 4595055325,
44
- "checkpoint": "checkpoints/4595055k/",
45
  "files": {
46
  "config.json": {
47
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4595055k/config.json",
48
- "sha256": "f1738b188c8db37652e2c4fb4888c39a80bb84b279cc804feac035143c333336",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4595055k/model.safetensors",
53
- "sha256": "a505413119b2284ef28d516b472a712f12188a71dff686a294482933b9afc8e0",
54
  "bytes": 32638048
55
  }
56
  }
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 122510,
4
  "quant": "ternary",
5
  "quant_scope": "all",
6
  "n_experts": 32,
 
40
  "pool_depth": 1,
41
  "pool_max": 32
42
  },
43
+ "tokens": 4599934752,
44
+ "checkpoint": "checkpoints/4599934k/",
45
  "files": {
46
  "config.json": {
47
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4599934k/config.json",
48
+ "sha256": "b212d8fb57404eeee31ec6fe1ab431a607e9d9076ff162a5110a0181168cc354",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4599934k/model.safetensors",
53
+ "sha256": "554cfa2eb9e7f702279f7c71e23ed19d795ec5a9c476f6234945e30bd43b123b",
54
  "bytes": 32638048
55
  }
56
  }
latest/config.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "name": "nirca-mini",
3
- "step": 122306,
4
- "tokens": 4595055325,
5
  "quant": "ternary",
6
  "quant_scope": "all",
7
  "n_experts": 32,
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 122510,
4
+ "tokens": 4599934752,
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:a505413119b2284ef28d516b472a712f12188a71dff686a294482933b9afc8e0
3
  size 32638048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:554cfa2eb9e7f702279f7c71e23ed19d795ec5a9c476f6234945e30bd43b123b
3
  size 32638048