Upload folder using huggingface_hub
Browse files- config.json +56 -0
- rank0.safetensors +3 -0
- rank1.safetensors +3 -0
- rank2.safetensors +3 -0
- rank3.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"mlp_bias": false,
|
| 3 |
+
"attn_bias": false,
|
| 4 |
+
"rotary_base": 500000.0,
|
| 5 |
+
"rotary_scaling": {
|
| 6 |
+
"factor": 8.0,
|
| 7 |
+
"low_freq_factor": 1.0,
|
| 8 |
+
"high_freq_factor": 4.0,
|
| 9 |
+
"original_max_position_embeddings": 8192,
|
| 10 |
+
"rope_type": "llama3"
|
| 11 |
+
},
|
| 12 |
+
"residual_mlp": false,
|
| 13 |
+
"disable_weight_only_quant_plugin": false,
|
| 14 |
+
"moe": {
|
| 15 |
+
"num_experts": 0,
|
| 16 |
+
"top_k": 0,
|
| 17 |
+
"normalization_mode": null,
|
| 18 |
+
"tp_mode": 0
|
| 19 |
+
},
|
| 20 |
+
"architecture": "LlamaForCausalLM",
|
| 21 |
+
"dtype": "bfloat16",
|
| 22 |
+
"vocab_size": 128256,
|
| 23 |
+
"hidden_size": 8192,
|
| 24 |
+
"num_hidden_layers": 80,
|
| 25 |
+
"num_attention_heads": 64,
|
| 26 |
+
"hidden_act": "silu",
|
| 27 |
+
"logits_dtype": "float32",
|
| 28 |
+
"norm_epsilon": 1e-05,
|
| 29 |
+
"position_embedding_type": "rope_gpt_neox",
|
| 30 |
+
"max_position_embeddings": 131072,
|
| 31 |
+
"num_key_value_heads": 8,
|
| 32 |
+
"intermediate_size": 28672,
|
| 33 |
+
"mapping": {
|
| 34 |
+
"world_size": 4,
|
| 35 |
+
"gpus_per_node": 8,
|
| 36 |
+
"tp_size": 4,
|
| 37 |
+
"pp_size": 1,
|
| 38 |
+
"moe_tp_size": 4,
|
| 39 |
+
"moe_ep_size": 1
|
| 40 |
+
},
|
| 41 |
+
"quantization": {
|
| 42 |
+
"quant_algo": null,
|
| 43 |
+
"kv_cache_quant_algo": null,
|
| 44 |
+
"group_size": 128,
|
| 45 |
+
"smoothquant_val": null,
|
| 46 |
+
"clamp_val": null,
|
| 47 |
+
"has_zero_point": false,
|
| 48 |
+
"pre_quant_scale": false,
|
| 49 |
+
"exclude_modules": null
|
| 50 |
+
},
|
| 51 |
+
"use_parallel_embedding": false,
|
| 52 |
+
"embedding_sharding_dim": 0,
|
| 53 |
+
"share_embedding_table": false,
|
| 54 |
+
"head_size": 128,
|
| 55 |
+
"qk_layernorm": false
|
| 56 |
+
}
|
rank0.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0cc28fe3bc8f03cee1807c9e6e907532d79af16e61f4e40c4d17894b2e61e0d2
|
| 3 |
+
size 47592378592
|
rank1.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:554a221554736f2434ea4f43dc8c88b291a1d1a1715e13bb0ba16102a3a8e93e
|
| 3 |
+
size 47592378592
|
rank2.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c4fd0f87efb997afce6ee8446d8082181515164bd92cc80fab1842f06ffaec93
|
| 3 |
+
size 47592378592
|
rank3.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4d0079cb73ddb908552a14bc5f314cc1120ee41f99d55fd2832a37211252fdd1
|
| 3 |
+
size 47592378592
|