tg-techie-agents commited on
Commit
250f0a5
·
verified ·
1 Parent(s): e74b391

latest.json: checkpoints/4527679k (step 119,617)

Browse files
Files changed (5) hide show
  1. README.md +4 -4
  2. latest.json +7 -7
  3. latest/config.json +2 -2
  4. latest/model.safetensors +1 -1
  5. lines.json +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
- 119,515 (4,525,031,361 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:** 119,515
119
- - **Tokens seen:** 4,525,031,361 (one token is one byte)
120
- - **Quantization-aware training:** from step 14,655 of 119,515, so for 99% of its 4,525,031,361 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
+ 119,617 (4,527,679,255 tokens seen).
34
 
35
  ## Model details
36
 
 
115
 
116
  ## Training procedure
117
 
118
+ - **Steps:** 119,617
119
+ - **Tokens seen:** 4,527,679,255 (one token is one byte)
120
+ - **Quantization-aware training:** from step 14,655 of 119,617, so for 99% of its 4,527,679,255 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": 119515,
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": 4525031361,
44
- "checkpoint": "checkpoints/4525031k/",
45
  "files": {
46
  "config.json": {
47
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4525031k/config.json",
48
- "sha256": "060efa108af61695bfd8d27930d7ba55d854c3a900ebbe036b8981d25fd332c9",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
- "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4525031k/model.safetensors",
53
- "sha256": "e512b7dc21db830eeb29a2345f5b4655ad0aec14461bf06a9f46d27570012529",
54
  "bytes": 32638048
55
  }
56
  }
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 119617,
4
  "quant": "ternary",
5
  "quant_scope": "all",
6
  "n_experts": 32,
 
40
  "pool_depth": 1,
41
  "pool_max": 32
42
  },
43
+ "tokens": 4527679255,
44
+ "checkpoint": "checkpoints/4527679k/",
45
  "files": {
46
  "config.json": {
47
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4527679k/config.json",
48
+ "sha256": "bbd76021c62d83e6c322a32a5bfcb1ee77b64340f37366b7fb5c13d2b9b500a2",
49
  "bytes": 874
50
  },
51
  "model.safetensors": {
52
+ "url": "https://huggingface.co/nirca/nirca-mini/resolve/main/checkpoints/4527679k/model.safetensors",
53
+ "sha256": "8754abc1579bcc6c55ee273bca35b7a2bcf34d64ded8bf1a017b4b94f02be80d",
54
  "bytes": 32638048
55
  }
56
  }
latest/config.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "name": "nirca-mini",
3
- "step": 119515,
4
- "tokens": 4525031361,
5
  "quant": "ternary",
6
  "quant_scope": "all",
7
  "n_experts": 32,
 
1
  {
2
  "name": "nirca-mini",
3
+ "step": 119617,
4
+ "tokens": 4527679255,
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:e512b7dc21db830eeb29a2345f5b4655ad0aec14461bf06a9f46d27570012529
3
  size 32638048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8754abc1579bcc6c55ee273bca35b7a2bcf34d64ded8bf1a017b4b94f02be80d
3
  size 32638048
lines.json CHANGED
@@ -11,7 +11,7 @@
11
  {
12
  "id": "mini-chat",
13
  "kind": "chat",
14
- "label": "Nirca Mini 0.1 (chat)",
15
  "channels": {
16
  "latest": "lines/mini-chat/latest.json"
17
  }
 
11
  {
12
  "id": "mini-chat",
13
  "kind": "chat",
14
+ "label": "Nirca Mini 0 (chat)",
15
  "channels": {
16
  "latest": "lines/mini-chat/latest.json"
17
  }