tg-techie-agents commited on
Commit
33840db
·
verified ·
1 Parent(s): dbfd902

latest.json: checkpoints/4573977k (step 121,424)

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
- 121,313 (4,571,327,597 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:** 121,313
119
- - **Tokens seen:** 4,571,327,597 (one token is one byte)
120
- - **Quantization-aware training:** from step 14,655 of 121,313, so for 99% of its 4,571,327,597 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
+ 121,424 (4,573,977,066 tokens seen).
34
 
35
  ## Model details
36
 
 
115
 
116
  ## Training procedure
117
 
118
+ - **Steps:** 121,424
119
+ - **Tokens seen:** 4,573,977,066 (one token is one byte)
120
+ - **Quantization-aware training:** from step 14,655 of 121,424, so for 99% of its 4,573,977,066 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": 121313,
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": 4571327597,
44
- "checkpoint": "checkpoints/4571327k/",
45
  "files": {
46
  "config.json": {
47
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4571327k/config.json",
48
- "sha256": "0b261f3c8ddf2c95b8a7b0aca690a20e9539a92c54667d966405f5f2dbb7c5aa",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4571327k/model.safetensors",
53
- "sha256": "495b6255498f26006fefe000c395f5f7ce2728ed5d1c6fadaeee4ec4fc8d7447",
54
  "bytes": 32638048
55
  }
56
  }
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 121424,
4
  "quant": "ternary",
5
  "quant_scope": "all",
6
  "n_experts": 32,
 
40
  "pool_depth": 1,
41
  "pool_max": 32
42
  },
43
+ "tokens": 4573977066,
44
+ "checkpoint": "checkpoints/4573977k/",
45
  "files": {
46
  "config.json": {
47
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4573977k/config.json",
48
+ "sha256": "91eaa92255d8a2bc56fd5cbcec7bf1dd631412e7a21a46bf71fc2e0c74477ad8",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4573977k/model.safetensors",
53
+ "sha256": "fddb1835914495b70ee95ef24dc18a3dd603d62418f79c6d605a4cd6ec7018c1",
54
  "bytes": 32638048
55
  }
56
  }
latest/config.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "name": "nirca-mini",
3
- "step": 121313,
4
- "tokens": 4571327597,
5
  "quant": "ternary",
6
  "quant_scope": "all",
7
  "n_experts": 32,
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 121424,
4
+ "tokens": 4573977066,
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 sha256:495b6255498f26006fefe000c395f5f7ce2728ed5d1c6fadaeee4ec4fc8d7447
3
  size 32638048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fddb1835914495b70ee95ef24dc18a3dd603d62418f79c6d605a4cd6ec7018c1
3
  size 32638048